financial stability report 2007 - cnb.cz · maintaining financial stability is defined as one of...

FINANCIAL STABILITY REPORT

20

07

FINANCIAL STABILITY REPORT

ISBN 978-80-87225-03-5

Maintaining financial stability is defined as one of the CNB's main objectives in Act No. 6/1993 Coll., onthe Czech National Bank, as amended:

Article 2

(2) In accordance with its primary objective, the Czech National Bank shall Ö..d) supervise the activities of entities operating on the financial market, analyse the evolution of

the financial system, see to the sound operation and development of the financial market inthe Czech Republic, and contribute to the stability of its financial system as a whole.

The CNB defines financial stability as a situation where the financial system operates with no seriousfailures or undesirable impacts on the present and future development of the economy as a whole, whileshowing a high degree of resilience to shocks.

The CNB's definition is based on the fact that financial stability may be disturbed both by processes insidethe financial sector leading to the emergence of weak spots, and by strong shocks, which may arise fromthe external environment, domestic macroeconomic developments, large debtors and creditors, economicpolicies or changes in the institutional environment.

The CNB's aim with regard to financial stability is above all to ensure a degree of resilience of the systemthat minimises the risk of financial instability. To fulfil this aim, the CNB as a monetary and supervisoryauthority uses the instruments made available to it by the Act on the CNB. Cooperation with other nationaland international authorities is also very important in this area. In order to maintain financial stability, theCNB focuses on prevention and broad communication with the public regarding the potential risks andfactors posing a threat to financial stability. This Financial Stability Report is an integral part of suchcommunication.

CONTENTS 5

LIST OF BOXES 6

LIST OF ALTERNATIVE SCENARIOS 6

SUMMARY 7

PART I 13

1 INTRODUCTION 14

2 THE REAL ECONOMY 162.1 THE MACROECONOMIC ENVIRONMENT 162.2 NON-FINANCIAL CORPORATIONS 232.3 HOUSEHOLDS 27

3 ASSET MARKETS AND THE FINANCIAL INFRASTRUCTURE 323.1 THE FINANCIAL MARKETS 323.2 THE PROPERTY MARKET 373.3 THE FINANCIAL INFRASTRUCTURE 42

4 THE FINANCIAL SECTOR 454.1 FINANCIAL SECTOR DEVELOPMENTS 46

4.1.1 The banking sector 464.1.2 Non-banking financial institutions 53

4.2 ASSESSMENT OF THE FINANCIAL SECTOR'S RESILIENCE 57

PART II 64

THEMATIC ARTICLES 64

THE ROLE OF RATINGS IN FINANCIAL SECTOR STABILITY ASSESSMENT 65SCORING AS AN INDICATOR OF FINANCIAL STABILITY 76COMPETITION AND EFFICIENCY IN THE CZECH BANKING SEKTOR 86OPERATIONAL RISK AND ITS IMPACTS ON FINANCIAL STABILITY 94

GLOSSARY 105

ABBREVIATIONS 109

SELECTED FINANCIAL STABILITY INDICATORS 112

6 LIST OF BOXES AND LIST OF ALTERNATIVE SCENARIOS

BOX 1: THE CREDIT CRUNCH IN THE USA AND ITS ROOTS 17BOX 2: ANALYSIS OF EXPORT-ORIENTED CORPORATIONS 25BOX 3: THE DEBT BURDEN ON HOUSEHOLDS AND LOAN REPAYMENT PROBLEMS 29BOX 4: FINANCIAL MARKET LIQUIDITY 34BOX 5: CARRY TRADES AND THE EXCHANGE RATE OF THE CZECH KORUNA 36BOX 6: IDENTIFYING PROPERTY MARKET BUBBLES 40BOX 7: STRESS TESTING OF BANKS' BALANCE SHEET LIQUIDITY 61

LIST OF ALTERNATIVE SCENARIOSALTERNATIVE SCENARIO A: "SAFE HAVEN" 21ALTERNATIVE SCENARIO B: "PROPERTY MARKET CRISIS" 25ALTERNATIVE SCENARIO C: "LOSS OF CONFIDENCE" 36

SUMMARY

From the financial stability point of view, the development of the Czech economyin 2007 can be assessed as very successful, notwithstanding the shocks emanatingfrom foreign financial markets. Despite a predicted moderation in economicgrowth, the current outlook for the next two years creates good conditions formaintaining the degree of financial stability already achieved. The main risks includea further deepening of the credit crisis in the advanced economies, a morepronounced slowdown in economic growth abroad and a continued very strongexchange rate of the koruna. These risks could slow net export growth, weakendomestic economic activity and, in turn, worsen the performance of the financialsector. In the context of the credit crisis abroad and the global increase in the riskaversion of financial institutions, some tightening of the lending conditions can alsobe expected in the Czech financial system, fostering slower credit growth. However,these effects can be viewed as a desirable stabilising reaction of the system, whichin the previous two years developed very dynamically and was affected byoptimistic expectations typical of the peak of the business cycle.

The successful development of the world economy continued into 2007. However,risks that had built up in the financial sector over the previous ten years started tomanifest themselves strongly in the USA and some other advanced countries. Theyshowed up initially in the US subprime mortgage crisis, which gradually spilled overinto other financial market segments and turned into credit crisis in the first fewmonths of 2008. This has led to a considerable downward revision of the outlookfor global economic growth for the coming two years. In addition to the USeconomy, which is highly likely to fall into recession, economic activity is likely toslow sharply in Japan and the UK. A downturn can also be expected in the euroarea. Yield curve slopes indicate substantial scepticism regarding a quick recovery ofthese economies.

The credit crisis poses substantial downside risks to economic growth and financialsystem stability in countries directly hit by the crisis, with potential knock-on effectson other economies. One of biggest risks over the next two years is a potentialcredit contraction in some advanced countries as problems with mortgage loansand related securities spread to other segments of the credit market. The majorlosses of some banks will lead to a decrease in the overall capital adequacy ofbanking sectors. The subsequent recapitalisation of banks will negatively affectgrowth in lending to the private sector and economic growth.

The turn in the credit cycle strongly affected the monetary policies of the keycentral banks. The most striking response came from the US Federal ReserveSystem, which between September 2007 and April 2008 lowered its monetarypolicy rate by 3.25 percentage points to 2%. The monetary policy tightening in theeuro area came to a halt, with the ECB holding its key rate at 4% over the entireperiod of turbulence. CNB monetary policy responded to rising inflationarypressures. The CNB increased its monetary policy rate four times during 2007 andonce again in February 2008 − each time by 0.25 percentage point. The CNB's May2008 macroeconomic forecast expects the one-off inflationary factors to startunwinding in 2008, with inflation gradually returning to the target. Accordingly,the interest rate environment in the Czech Republic should be stable over the nexttwo years.

The downward revision of the outlook for interest rates of the main worldcurrencies led to renewed interest in investing in stable emerging markets' assets.Given the perception of the Czech koruna as a "safe haven", the risk of renewedappreciation pressures on the koruna materialised. The koruna appreciateddramatically in the initial phase of the financial turbulence, probably due to theliquidation of carry trades for which the koruna was used as the financing currency,in an environment of growing risk aversion. The increasing appreciation trend in theperiod that followed was driven both by a shift in foreign investment away from the

7

The world economy recorded rapideconomic growth, but this will weakenconsiderably over the next two yearsowing to the credit crisis in someadvanced countries

Owing to banks' losses, countries hit by the crisis will experience a sharpslowdown in growth in lending to the private sector

The turn in the credit cycle promptedmixed reactions from central banks

The Czech koruna appreciated stronglyin response to the financial turbulence

depreciating dollar and towards appreciating currencies such as the Czech koruna,and by domestic exporters, who in an environment of faster appreciation of thekoruna started hedging on a mass scale by selling euros. The excessively fastappreciation of the koruna combined with falling external demand as a result of theeconomic slowdown in the euro area could have a negative effect on the domesticeconomy, primarily through a deterioration in net export growth. Given this risk, theCNB signed an agreement with the Czech Government in April 2008 on measuresto prevent public sector operations from having undesirable impacts on the foreignexchange market and subsequently on macroeconomic stability.

The domestic macroeconomic situation and financial sector were affected by the approved reform of public finances, whose expected impacts started affectingthe behaviour of economic agents as early as 2007. The fiscal measures adoptedare having conflicting impacts on the disposable income of households andcorporations. The changes in the fiscal impulse and the impacts of the expectedchanges on the behaviour of economic agents will cause some volatility inaggregate demand. Efforts to buy real estate still at the lower VAT rate gave rise toincreased demand for loans for house purchase and also affected the constructionindustry's performance. The opposite effect can be expected in the coming years,probably compounded by a review of the real estate market risks by banks inresponse to the problems of this market in advanced economies.

At the end of 2007, despite the continued robust growth of the domestic economy,clear signals of a future slowdown were observable, mainly as a result of slowergrowth in real disposable income due to higher inflation, subdued investmentactivity and slowing net exports. The CNB's May 2008 macroeconomic forecastexpects real GDP growth to slow to 4.7% in 2008 and 4.0% in 2009. The slowingeconomic growth and declining growth of real disposable income may have someimpacts on the ability of households and corporations to repay loans taken out inprevious years. On the other hand, it may help to eliminate the excessivelyoptimistic expectations which emerged at the peak of the cycle and supported arapid rise in the indebtedness of households and some segments of the non-financial corporations sector.

The traditional macroeconomic sustainability indicators recorded strongly positivedevelopments in 2007. The public budget deficit under ESA95 methodology fell to1.6% of GDP and the ratio of public debt to GDP decreased to 28.7%. The currentaccount deficit declined to 2.5% of GDP, while the surplus on the output balanceincreased.

Non-financial corporations posted good financial results in 2007, as evidenced byimproved profitability indicators, liquidity, value added per employee and inventoryratios. Although the corporate debt ratio continued rising in 2007, the growth rate of loans granted to non-financial corporations decreased. This may indicate afuture downswing in corporate sector performance associated with the expectedweakening of economic activity. The credit risk of the corporate sector as measuredby the 12-month default rate remained below the 3% level during 2007. However,given the expected economic slowdown, the macroeconomic credit risk modelassumes a rise in this indicator of 1−2 percentage points during 2008. A slightincrease in the corporate sector's credit risk in 2008 is also indicated by the newcreditworthiness indicator.

The economic condition of non-financial corporations started to be affectedsignificantly in the second half of 2007 by the sharp appreciation of the koruna.Although the first-quarter data on the real economy in 2008 suggested that export-oriented corporations were able to cope with this exchange rate shock, somenegative effect can be expected to emerge gradually. This will probably be reflectedin an increase in the sector's credit risk, as indicated by an analysis of monthly data

8 SUMMARY

The domestic economy started to beaffected significantly by the fiscal

reform

Czech GDP growth will slow over thenext two years

Macroeconomic sustainability indicatorsrecorded a noticeable improvement

The good financial condition ofcorporations fostered low credit risk

The sharp appreciation of the korunawill be reflected in a rise in the default

rate of export-oriented corporations

SUMMARY

on the repayment of loans provided by banks to the largest exporters in the CzechRepublic. The analysis revealed that the 12-month default rate is more volatile forexporters than for the economy as a whole, so it can be expected to deviateupwards at a time of strong appreciation. The analysis also revealed that for part ofexporting corporations borrowing in foreign currency is a significant way ofreducing their sensitivity to appreciation of the koruna.

Household debt increased considerably in 2007. As in previous years, the overalldebt of households to financial corporations increased by roughly one-third,breaking through the CZK 800 billion level. Loans to households reached CZK 680 billion in banking institutions and CZK 150 billion in non-bankinginstitutions at the close of the year. The growing debt was supported mainly byloans for house purchase, as a result of demand for owner-occupied housingbolstered by the one-off effect of advance financing of construction work owing toa change in the lower VAT rate from 5% to 9% with effect from 1 January 2008.The increasing level of debt is being accompanied by a decreasing gross saving rateof households (5.1% in 2007), which fell to a ten-year low. However, despite therapid credit growth, the overall Czech household debt level is still low bycomparison with the Western European countries and, according to foreignanalyses, is at a safe level. In 2008, the rate of growth of household debt shouldstart decreasing, as confirmed by the first-quarter growth in new loans for housepurchase.

Although Czech households as a whole cannot currently be regarded asoverleveraged, some risk of growth in their default rate exists if the less favourablemacroeconomic scenarios materialise. One potential source of overindebtedness isgrowth in mortgage repayments, associated with the purchase of ever moreexpensive property. Owners of loans for house purchase whose income has beenrising nowhere near as fast as property prices are particularly exposed to this risk.This overindebtedness may come out into the open with a lag, for example as aresult of a rise in interest rates during the refixation of a mortgage loan. Consumercredit is another possible source of overindebtedness, particularly for low-incomehouseholds. A sharp rise in the number of executions ordered suggests that this riskis growing over time.

The shocks on the advanced financial markets manifested themselves in decliningprices of a whole range of risky assets and rising volatility on equity, bond andforeign exchange markets. However, the domestic interbank money marketremained fully operational and interest rates there were affected mostly byexpected changes in CNB monetary policy. The Czech financial markets respondedto developments in global financial markets similarly as during corrections inprevious years. Share prices went down in line with the falls in foreign stockmarkets, while bond yields initially declined slightly and then stabilised.Nevertheless, it cannot be said that the global financial crisis had no major effecton the Czech financial markets. During 2008 Q1, market signals indicatedincreased sales of Czech government bonds by foreign investors, fostering growthin the risk premium of the Czech koruna and the euro government debt. Theinterbank market also recorded a very modest increase in the credit risk premiumand a decline in market liquidity. The Czech banking sector responded to the risein global risk aversion only with a very moderate tightening of the interest rateconditions for loans for house purchase and some riskier segments of the creditmarket.

The property market trends identified as risky in the 2006 Financial Stability Reportcontinued into 2007. Property prices in the Czech Republic rose relatively fast in2007 despite the problems on real estate markets in many advanced economies.Prices of flats and building land recorded high growth, while prices of family housesrose more slowly. Although the property market growth still cannot be regarded as

9

Despite rapid credit growth, householddebt is still low by internationalcomparison

In the event of adverse macroeconomicdevelopments, overindebtedhouseholds − particularly those fromlow-income groups − will get intodifficulties

Czech financial markets were affectedonly slightly by the external shocks

Property prices continued rising rapidlyin 2007, while rent returns fell further

a bubble and the rise in prices as manifestly unbalanced, there are some indicationsthat the market is overheating. For example, the ratio of property prices tohousehold income is rising over time. This indicator identifies Prague as the riskiestregion. A large rise in supply prices of flats together with considerably lowerincreases in supply rents led to a further fall in the rent return, which for most largecities in the Czech Republic is below the level of long-term bond yields and the levelof interest rates on new loans for house purchase. Speculative property purchasesfinanced by mortgage loans thus become less profitable and more risky.

The share of property developers in total loans to the corporate sector increased to more than 25% in 2007, compared to less than 10% in 2001, against abackground of rising prices of both residential and commercial property. Loansprovided to real estate companies showed record growth of 50% in 2007. Thedevelopments in 2007 confirmed that developers tend to react with a lag to growthin property prices. Before the Czech Republic joined the EU, strong growth inproperty prices in 2002−2004 gave rise to overly optimistic expectations, which ledto higher investment activity by developers. The subsequent sharp slowdown inproperty prices then brought about a rise in the rate of default by developers in2005−2006. The property development sector responded with increased activity tothe property price growth in 2006 and 2007 again with a lag, so it is highly likelythat a negative shock would be reflected in a significant rise in the previouslyrelatively low default rate in this sector.

The number of housing completions increased by an exceptional 38% year on yearin 2007. It cannot be ruled out that the market for new flats will become saturatedand that the rise in prices will slow down or stop as a result of the rising supply. Anincrease in the difference between supply prices and the prices of final transactionsmay be a signal of decreasing market liquidity or rising property market risk. Acooling of the property market represents quite a sizeable risk for the bankingsector. According to stress tests, the banking sector is resilient to this risk at present.In response to the property market problems in some advanced countries, domesticbanks considerably tightened their credit standards vis-a-vis the propertydevelopment sector in early 2008. This can be viewed as an appropriate responseto the evolution of the risks in the sector.

The financial infrastructure continued running reliably in 2007. The CERTISinterbank payment system and the SKD short-term bond settlement systemrecorded no irregular situations in the period of turbulence in global financialmarkets. Smooth and stable interbank settlement was aided by a gradual rise in theuse of intraday credit by CERTIS participants. The fact that domestic banks did notneed to use intraday credit any more often than in the past, despite the liquidityproblems on foreign financial markets, is meanwhile evidence that the financialcrisis had very small direct impacts on the Czech financial sector. One of theinfrastructural shortcomings of the Czech capital market is the absence of a centraldepository.

In 2007, as in previous years, the Czech financial sector experienced mostly positivedevelopments. Banks and insurance companies achieved high returns on assets andequity. Provided that a reasonable proportion of the profit remains in financialinstitutions in the form of equity capital, this lays the groundwork for maintaininga high level of financial stability in the years to come. The depth of financialintermediation in the Czech Republic, as measured by the ratio of financial sectorassets to GDP, increased by just under 10 percentage points to 142%. This relativelylarge increase reflects a rise of almost 18% in financial institutions' assets, duemainly to growth in bank loans of more than 26%.

The available analyses indicate that the Czech banking sector is not exposed to therisk of a crisis similar to the one that hit the US subprime mortgage segment. This

10 SUMMARY

In response to property marketdevelopments, the banking sector

tightened its credit standards vis-a-visthe property development sector

The financial infrastructure continuedrunning reliably in 2007, but the

further development of the Czechcapital market is being hampered by

the absence of a central depository

The depth of financial intermediationincreased significantly, unlike in the

previous year

The characteristics of the Czechhousing loan market minimise the risk

of a crisis similar to the one that hit theUS subprime mortgage segment

The risk of default by real estatebusinesses depends to a large extent

on the future evolution of propertyprices

SUMMARY

is due to higher required debtor creditworthiness, the traditional method of interestrate fixation, less use of external mortgage underwriters, the absence of creditsecuritisation and, in particular, good collateralisation of mortgage loans withproperty. Mortgage loans, i.e. loans that are fully secured by property, account for65% of total loans for house purchase. The mortgage loan to property value ratio(the LTV ratio), which for mortgage loans to households was 56% on average atthe end of 2007, can be still regarded as conservative. For the same reasons, theratio of loans in default to mortgage loans was only 1.2% at the end of 2007. The ratio of loans in default to total loans stood at 2.7%, falling by almost 1 percentage point year on year. Given the favourable phase of the business cycleand the high credit growth in 2007, the loan default rate still probably undervaluesthe magnitude of the banking portfolio credit risk to some degree.

Client deposits remain the biggest source of financing for bank loans. At the endof 2007, they were 1.3 times higher than client loans, which, in turn, is more thantwo times the average in the original EU member countries. The large volume ofclient deposits provides domestic banks with protection against any rapid drying upof market liquidity and at the same time ensures stable and relatively low-costfunds compared to other forms of external financing. However, deposit growth hasbeen lower than credit growth in the Czech Republic for several years and this trendis likely to continue. As a result, the share of deposits in bank funds will decreasein the future and banks will have to respond with changes in balance-sheet liquiditymanagement. Tests of banks' balance-sheet liquidity indicate that the bankingsector is resilient enough to deposit outflows and some other hypothetical changesin the financial market. However, for the extreme variant of pressures on balance-sheet liquidity, only institutions with a strong deposit base are naturally resilient.

The preparations for the implementation of Basel II and the actual changeover tothe new prudential rules in several banks on 1 July 2007 were a significantchallenge for the banking sector. The remainder of the sector took this step inJanuary 2008. Owing to the switch to Basel II, there was a slight decline in theregulatory capital requirements. The CNB's analyses predict a slight rise in defaultrates for corporations and households in 2008. This would imply some increase inthe regulatory capital requirements in the period ahead.

In the insurance and pension scheme industries, the previous years' trendscontinued into 2007. Premiums written grew at an increased rate of 9% year onyear, mainly due to growth in life insurance. Insurance companies met the solvencycriteria − the aggregate available solvency margin under the current legislation was3.0 times the required solvency margin on the life insurance market and 3.3 timesthat on the non-life insurance market. Contributions from pension planholdersincreased by more than 14% in 2007. Insurance companies and pension funds sawmarked growth in the costs of intermediating new contracts. The growth in thesecosts may encumber the private pension system in particular. Growing volatility onthe asset market and potential asset revaluation losses in particular may have anadverse effect on the funds' assets and performance, particularly in the event ofhigher and faster payments of benefits. Commensurate capital increases by funds'shareholders would be desirable as protection against the existing risks.

According to stress tests, the financial sector is currently resilient to the market,credit and some specific risks to which it is exposed. However, an extrememacroeconomic scenario with significant adverse impacts on interest rates, theexchange rate and GDP growth would necessitate capital injections to ensurecompliance with the regulatory limits and maintain sufficient capital adequacy infinancial institutions. The aggregate banking sector stability indicator confirms acontinuing process of capital optimisation in the banking sector, with unchangedresilience to the main risks.

11

The large volume of client depositsprovides domestic banks withprotection against any drying up ofmarket liquidity and ensures stableand low-cost sources of financing forcredit expansion

Owing to the gradual changeover toBasel II, capital adequacy increasedslightly Ö but an expected rise inhousehold and corporate default rateswill lead to an increase in theregulatory capital requirements

The insurance and pension schemeindustries continued growing in 2007,although higher asset price volatilityand growth in the costs ofintermediating new contracts arecreating risks for pension funds inparticular

According to stress tests, the financialsector seems to be resilient to a widerange of risks. Only an extrememacroeconomic scenario wouldnecessitate capital injections tomaintain sufficient capitalisation

PART I

1 INTRODUCTION

The Czech National Bank is pleased to present to the public its fourth FinancialStability Report, this time for 2007. This Report analyses the risks to the financialstability of the Czech Republic in the near future on the basis of previous andexpected developments in the real and financial sectors. The turbulence on globalfinancial markets that emerged in summer 2007 highlights the relevance of theseanalyses, especially as regards the potential impacts of external developments onthe Czech financial system.

Starting with this Report, the work with the risks to financial stability is shiftingtowards a more detailed analysis of selected risks and quantification of theirimpacts on the stability of the financial system. This shift allows us to enhance theconsistency of the analyses and raise the Report's level towards the highestinternational standards. This year's Report profits from a gradually constructedmodelling and analytical framework for the financial stability area based onadvanced stress testing and accompanying economic models. Three alternativescenarios of adverse developments were constructed on the basis of an analysis oftrends and weak spots in the domestic and external economy and financial sector,and their impact on the financial sector was tested. The alternative scenarios takeinto account the current turbulence on global financial markets, developments inthe Czech property market and the risks of a potential cooling of economic activityabroad. The baseline scenario from the CNB's official February 2008macroeconomic forecast is used for comparison. The alternative scenarios − entitled"safe haven", "property market crisis" and "loss of confidence" are presented inthe Report in the form of boxes in the sections that analyse the main componentsof the individual scenarios. A new subsection 4.2 Assessment of the financialsector's resilience contains a discussion of the impacts of all the scenarios.

Our efforts to strengthen the quantitative aspects of the Report necessitated areview of the interactions and transmission channels for the spread of instabilityacross individual economic sectors. These linkages are quantified using sub-models,expert estimates and other analytical techniques. The basic analytical tool foranalysing the spread of instability and assessing the impacts of shocks to financialsystem stability is stress testing, which the CNB has been developing for severalyears now. The development of the stress-testing methodology and the resultsderived from the various scenarios for development have been regularly disclosedby the CNB in the form of annexes and thematic articles in previous Reports. Thestress tests quantify the impact of changes in key macroeconomic and financialindicators on the solvency of the most important financial institutions, i.e. banks,insurance companies and pension funds. These tests are complemented by severalquantitative indicators of the stability of the financial system, such as the bankingsector stability indicator presented as a thematic article in last year's Report and thenon-financial corporation creditworthiness indicator described in a thematic articlein this Report.

This Report is similar in structure to last year's. The section entitled The realeconomy discusses developments in the external and domestic macroeconomicenvironment and in the key domestic sectors, i.e. households and corporations. Thesection entitled Asset markets and the financial infrastructure analyses the financialmarkets, the property market and the financial infrastructure. The last section, Thefinancial sector, covers developments in the financial sector and newly also theaforementioned section assessing the Czech financial market's resilience to shocksfrom the three alternative scenarios. The Report ends with a table of key indicatorsrelevant from the point of view of financial stability.

Like last year, the Report includes four thematic articles. The Role of Ratings inFinancial Sector Stability Assessment discusses the ratings of financial market

14 1 INTRODUCTION

1 INTRODUCTION

participants as one of the indicators of financial stability. Scoring as an Indicator of Financial Stability deals with the construction of an indicator of thecreditworthiness of non-financial corporations using a scoring model and is basedon a microeconomic study of Czech corporations. Competition and Efficiency in theCzech Banking Sector offers empirical evidence on the evolution of competition inthe Czech banking sector and its relationship to the cost efficiency of banks in theCzech Republic. The last article, Operational Risk and its Impacts on FinancialStability, describes the nature and significance of operational risk with regard tofinancial stability and identifies the impacts of the newly introduced capitalregulation of operational risk.

This Financial Stability Report was approved by the Bank Board of the Czech NationalBank on 15 May 2008. It is available in electronic form at http://www.cnb.cz/.

15

1 In its World Economic Outlook (April 2008), the IMF expects weaker growth of the global economy over the next two years, at 3.7%. The IMF's forecast is significantly more pessimistic than the average forecast in Consensus Forecast (a publication presenting the average estimates of a broadrepresentative sample of analysts and forecasters). Chart II.2 also illustrates how sharply the IMF revised its forecasts for 2008 between January 2008 (2008fS07) and April 2008 (2008f).

2 THE REAL ECONOMY

2.1 THE MACROECONOMIC ENVIRONMENT

External and domestic macroeconomic developments remained favourable in 2007.However, risks that had built up in the foreign financial system over the previousten years started to manifest themselves strongly in the second half of the year. Thefinancial turbulence, which was seen first in the USA in August, spilled over intosome other advanced economies and turned into credit crisis in the first fewmonths of 2008. These events started a sharp turn in the credit cycle, giving rise tothe risk of a credit contraction in countries directly hit by the crisis and a subsequentdownturn in economic activity. The turn in the credit cycle is also manifesting itselfin considerable volatility in asset prices, interest rates and exchange rates. Thissituation poses substantial downside risks to economic growth and financial systemstability in some advanced economies. At the same time it is generating a high levelof uncertainty, fundamentally hampering the forecasting of future macroeconomicdevelopments. The next two years can therefore be regarded as a period of largerisks.

The Czech Republic has not been hit directly by the financial crisis so far, and thedomestic financial sector is showing a high degree of resilience. A major side effectof the financial turbulence is a sharply appreciating koruna. Although the first-quarter data from the real economy indicated that corporations were able to copewith this exchange rate shock, some negative effect on exporters can be expectedto emerge gradually. Despite the expected slowdown in economic growth, thecurrent outlook for the Czech economy for the next two years can be regarded aspositive, albeit with sizeable risks originating particularly from abroad. A furtherdeepening of the credit crisis in the advanced economies and a continued strongor even stronger exchange rate of the koruna would have an adverse effect on thedomestic economy, in particular through deteriorating net export growth.

The successful development of the world economy, which, according to IMFmethodology, grew by 5% as a whole (the same as in 2006), continued into 2007.However, the record-high economic activity will weaken over the next two years,mainly as a result of falling economic growth in the advanced countries1 (see Chart II.1). Risks that had built up in those countries over the previous ten yearsstarted to manifest themselves strongly in 2007 H2. They showed up initially in theUS subprime mortgage market crisis (see Box 1) and started to spill over into othereconomies and various financial market segments.

The economic growth outlook was gradually reviewed downwards (see Chart II.2)in response to the deepening problems in financial systems. In addition to the USeconomy, which is highly likely to fall into recession, economic activity is likely toslow significantly in Japan and the UK (see Chart II.1). A downturn can also beexpected in the euro area, although this will not necessarily be as sharp. Theuncertainty regarding financial systems going forward meant that the uncertaintyregarding expected economic growth increased as well (see Chart II.3).

16 2 THE REAL ECONOMY

CHART II.1Economic growth in advanced economies (year-on-year growth in %; outturns and September 2007 and April 2008 forecasts)

0

1

2

3

4

5

World Advancedeconomies

Euroarea

USA Japan UK

2006 2007 2008fS07 2008f 2009f

Source: IMF (World Economic Outlook, April 2008)

CHART II.2Expected and actual economic growth in the USA andthe euro area(quarterly data; year-on-year growth; outturns versus expectations of ConsensusForecast, CF)

GDP growth in USA (in %; left-hand scale)GDP growth in euro area (in %; right-hand scale)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

2006 2007 2008 20090.511.522.533.544.555.5

Source: Eurostat, US Bureau of Economic Statistics, Consensus Forecast

CHART II.3Uncertainty regarding expected GDP growth in selectedeconomies(mean standard deviation of GDP growth estimates for given and next year fromConsensus Forecast; in percentage points)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

01/06 04 07 10 01/07 04 07 10 01/08

USA Euro area CZ

Source: Eurostat, US Bureau of Economic Statistics, Consensus Forecast

2 THE REAL ECONOMY 17

2 The specific terms and abbreviations contained in this Box are explained in the glossary. 3 The major global banks saw particularly large increases in leveraging (in some cases reaching 30 or

more), i.e. their total assets grew faster than their regulatory capital. This was because their newly generated assets had very low risk weights. At the same time, they recorded a sharp declines in theirdeposit-to-loan ratios, as it was very easy to acquire sufficient liquidity from the money markets.

4 The lending process changed considerably after assets with BBB or lower ratings started to be transformed into assets with AAA ratings (see section 3.1) through CDOs (collateralised debt obligations; see the glossary). Specifically, subprime mortgages started to be converted into CDOs,within which they were divided into tranches with various risk levels and some acquired an investmentrating. Any losses from the set of underlying mortgage loans were to be borne by the holders of the riskier lower-rated tranches. The low risk of the best tranches was conditional on a low default correlation for the underlying subprime mortgages, which, given their similar characteristics and theadverse development of the economic environment, proved to be illusory.

Box 1: The credit crisis in the USA and its roots 2

The roots of the current credit crisis in the USA stretch back to the 1980s and1990s. Economic growth was rising in the context of ongoing globalisation,and nominal interest rates were declining considerably thanks to fallinginflation. In this environment, corporate earnings and share prices wererising. The positive sentiment gradually fed through to prices of other typesof assets, in particular property prices. Underlying these developments weresignificant changes in the financial sector, which innovated and increased itsprofitability and its share in the economy. The dynamic growth was supportedby the Fed's accommodative monetary policy, which at times of heightenedrisks responded by easing the monetary conditions (particularly after 11September 2001). The favourable environment for the development offinancial activities was also supported by a rising supply of free savings fromthe emerging economies on international financial markets. This resulted inan extraordinary increase in free liquidity available to a wide range ofinvestors at a low price. At the same time, a "shadow" (parallel) bankingsystem started to develop vigorously, within which credit creation wassupported by a network made up of structured investment vehicles (SIVs,conduits), investment banks, investment funds, hedge funds and monolineinsurers. The shadow banking sector partially reduced the importance oftraditional banking intermediation (de-intermediation) and fostered a markedincrease in leveraging in the financial and non-financial sectors.3 This started happening in particular at the end of the 1990s, when, in addition to federal agencies, new private players providing subprime mortgages to"problem" debtors with higher probabilities of default started entering themarket. Subsequently, loans structured into various risk segments by meansof complex financial instruments began to be provided. These segments(tranches) were then sold separately to investors.4 The development of suchinstruments enabled the increased risk associated with the new products andthe growing leveraging to be diversified, to a certain extent masked andpartially transferred to insufficiently informed investors.

The expansion of the shadow banking system was driven by securitisation,which facilitated the development of an "originate-and-distribute" model. By means of securitisation, banks transferred the loans they had providedoutside their balance sheets, thanks to which they were able to increase theirregulatory capital much more slowly. Securitisation never involved justtransferring assets outside bank balance sheets. For it to work, certainconditions needed to be met. Banks transferring assets outside their balance


5 The relationship between low short-term interest rates and the rise in leveraging in the US banking sector is commented on by, for example, The Economist of 22 March 2008 in an article entitled "Thefinancial system: What went wrong?".

sheets lent to entities that invested in assets generated by securitisation andalso to entities that then bought derivatives derived from assets generated bysecuritisation. The bulk of the risks were diversified in this system (at leasttheoretically). However, it was necessary to seek final asset holders willing tobear the outstanding risk. This search was largely successful, since the riskswere partially hidden by the ratings of rating agencies, which were not wellprepared for valuing some products of the shadow financial system.

The changes in the nature of the mortgage market were reflected in poormonitoring of lending processes and a substantial relaxation of creditstandards. This relaxation took the form of new types of mortgages creatingan illusion of easy future repayment (lax or non-existent checking ofapplicants' creditworthiness, loans for 100% or more of property purchaseprices, instalments initially lower than interest, low introductory rates, fixedrates for the first few years and then a switch to floating rates). Manyinappropriate incentives permeated the originate-and-distribute model. Theentities involved along the entire chain were primarily interested in receivingfees for their share in the transaction and then transferring the risk to anotherholder. Mortgage brokers were interested in maximising the number andvolume of contracts, assessors were interested in revaluing property,mortgage loan originators (banks) transforming mortgages into collateralisedsecurities and banks packaging these securities into structured creditderivatives were interested in maximising the volume of loans and derivedsecurities, and rating agencies were interested in maximising the number ofratings. The advanced securitisation of mortgages into CDO-type productsenabled lending to the subprime segment to expand and credit standards tobe relaxed. By the middle of the decade, the relaxation of credit standardshad acquired extreme dimensions. This contributed to the subsequent failureof subprime-mortgage-backed CDOs. Prices of AAA tranches of mortgage-backed CDOs provided later on (particularly in 2007 H1) declined more andfaster than tranches with mortgages dating from 2006, as the underlyingmortgages provided later on were defaulting much faster because they hadprobably been provided to very risky debtors (see Chart II.1 Box).

At the end of this chain, of course, stood investors willing to accept credit risk. This would probably not have worked in a normal yield environment.However, a low-interest-rate environment had prevailed for many years,despite the relatively strong economic activity, and so investors with"surplus" funds seeking higher yields on the markets were accepting highlyrisky assets. Monetary policy also probably contributed to the build-up of risksin the financial sector. The expansion of the shadow banking system wasdependent to some extent on cheap short-term financing, since its mode ofoperation consists largely in investment in long-term assets financed withshort-term funds. And it was low short-term interest rates that, especiallybetween 2001 and 2005, supported the smooth operation of the shadowbanking system and the expansion of activities in the corporate sector withhigh leveraging.5 The relationship between monetary policy, the shadowbanking system and the credit crisis will undoubtedly be a focus ofmacroeconomic and financial research in the years to come.

CHART II.1 (BOX)Prices of AAA bonds backed by subprime mortgagesissued in individual half-years(ABX-HE index; prices in % of nominal value of underlying bonds)

50

60

70

80

90

100

1/07 4 7 10 1/08 4

2006-1 2006-2 2007-1 2007-2

Source: JP Morgan Chase

2 THE REAL ECONOMY

A major indicator of the size of the risks is the fact that at the end of 2007 long-term interest rates in the USA and the euro area fell below short-term rates and theyield curve thus became inverted at longer maturities (see Chart II.4). The yieldcurve in the USA has failed to turn significantly positive even after the sharp cuts inmonetary policy rates in recent months. This can be regarded as an indicator of astrong lack of confidence in a quick economic recovery. The yield curve in the CzechRepublic has also become flatter, but its slope is still normal.

The uncertainty regarding future economic growth in the advanced economies islinked chiefly with uncertainties regarding financial institutions' total losses (seesection 4). In its Global Financial Stability Report (April 2008), the IMF estimates thatinternational banks will suffer a loss of up to USD 500 billion, which will lead to a fall in total capital adequacy of 2.5 percentage points in the USA and 1.5percentage points in Europe. The need to recapitalise banks coupled with thelimited functionality of the securitisation process will cause a contraction (or at leasta marked slowdown in growth) in loans to the private sector.6 The IMF predicts thatthis will cause annual GDP growth in the USA to decline by 0.8−1.4 percentagepoints. A similar impact can be expected on some European economies, especiallythose in which property prices had been rising fast in previous years.

The impact of the potential US recession on global economic activity should besmaller than in similar situations in the past, thanks mainly to the fact that manyemerging economies, which export only a relatively small proportion of their outputto the USA, are in a phase of recovery or dynamic growth. The bulk of their exportsgo to other emerging economies or advanced European countries. The currentforecasts foresee continuing high growth of Chinese economy and satisfactorygrowth in the new EU Member States and the CIS countries (see Chart II.5).

19

6 At present, the global financial system is in a phase of return to a more traditional structure via processes referred to as de-leveraging and re-intermediation. The first term denotes processes leadingto a decline in leveraging in the economy as a whole. They involve a reduction in the availability of credit in the traditional and shadow banking systems, especially for financing riskier clients, and investment projects. The second term describes a re-strengthening of the position of asset-side and liability-side transactions across the entire financial system. The process also involves redirecting some activities from the shadow banking system back to regulated banks. In the initial phase, theseinterconnected processes may lead to some credit contraction.

CHART II.4Difference between long-term and short-term rates(%)

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

3/05 6 9 12 123/06 3/07 3/086 9 6 912

CZ EUR USSource: Datastream

CHART II.5Economic growth in emerging and developing countries (year-on-year growth in %; outturns and September 2007 and April 2008 forecasts)

0

2

46

8

10

12

Emergingand

developingeconomies

Centraland

EasternEurope

CIS China India CzechRepublic

2006 2007 2008fS07 2008f 2009f


CHART II.6Effective exchange rates of the euro(January 2001=100)

801/01 11 9/02 7/03 5/04 3/05 1/06 11 9/07

90100110120130140

NEER22 NEER3Asia REER22CPI

Souce: ECB, CNB calculationNote: The 22 most important trading partners of the euro area / Japan, China and Koreafor 3Asia.

To sum up, owing to the close links between participants in the shadowbanking system and banks, the transfer of risks to final investors viasecuritisation proved to be illusory to some extent. This was fully revealedwhen the prices of the assets generated in this system started falling.Potential investors dried up, banks were no longer able to provide credit forpurchases of securitised assets and the system practically ground to a halt. Ithas also been confirmed that leveraging is strongly pro-cyclical. When themarket went up, debt grew faster than assets. Financial groups were buyingsecurities on credit, and those securities were then being used as collateral forthat and other credit. The possibility of borrowing cheaply thus automaticallyfostered growth in asset prices and vice versa. As soon as the markets turned,the system accelerated again in the opposite direction. In a falling market,financial institutions had to sell their assets in order to repay their obligations.This reduced securities prices, worsened balance sheets and forced furthersales. This is how the credit crisis was able to take hold fully.

The risks in the world economy are also being exacerbated by persisting globalimbalances, i.e. different current account developments in the key economies (seeChart II.11). Thanks to the weak dollar, the US deficit is gradually decreasing. Asiancountries are also trying to maintain their currencies at relatively weak levels bymeans of intervention purchases of dollars (over USD 600 billion in 2007) andthereby support their high trade surpluses. Under this pressure, the euro area,which has long been balanced but whose currency has been steadily appreciatingin recent years (see Chart II.6), is getting into a slight deficit.

The turn in the credit cycle strongly affected central bank monetary policies (seeChart II.7). After a phase of monetary policy tightening by the Fed, which ended inJune 2006 after it raised its monetary policy rate to 5.25%, US interest rates stayedflat until September 2007. In 2007 H2, with the outlook for the US economyworsening, expectations of cuts in monetary policy interest rates in the USA startedto dominate. These expectations were subsequently realised by a series of cuts (twoof them by as much as 0.75 percentage point). At the end of April 2008, themonetary policy interest rate was 2.0%. Given the persisting inflationary pressures,the future monetary policy steps of the Fed and some other central banks areuncertain.7

In the euro area, the ECB raised monetary policy interest rates in two steps by 0.25percentage point in 2007. Since June 2007, rates have stayed at 4%. Expectationsof a halt in growth − or a decrease − in monetary policy interest rates in the euroarea started emerging in early 2008, amid a sharper slowdown in economic activity(see section 3.1). However, these expectations were suppressed by persistingrelatively high euro-area inflation, which reached 3.5% in March 2008. In theCzech Republic, the monetary policy rate was increased four times during 2007 andonce again in February 2008 − each time by 0.25 percentage point − in responseto the outlook for inflationary pressures. The CNB's May 2008 macroeconomicforecast assumes that the interest rate environment in the Czech Republic will berelatively stable over the next two years.

The downward revision of the global interest rate outlook led to renewed interestin investing in stable emerging economies' assets. A "search for yield" by investors,seen particularly in the first half of the decade (see the 2006 Financial StabilityReport), thus started to re-emerge. In this situation, the risk of renewedappreciation pressures on the koruna materialised. Like the other Central Europeancurrencies, the koruna has offered foreign investors relatively high returns in manyprevious years (see Chart II.8). The perception of the Czech koruna as a "safehaven" is being reinforced by its long-running steady nominal appreciation trend.The koruna started appreciating at the end of summer, probably due to theliquidation of carry trade positions (see section 3.1). In 2008 Q1, the appreciationbecame exceptionally strong, probably because of the safe-haven effect. Betweenthe onset of the financial turbulence in August 2007 and the start of April 2008,the koruna appreciated by 12% against the euro and by 25% against the dollar. Itappreciated against all traded currencies in this period (see Chart II.98).


7 The Fed's dramatic monetary policy measures can be explained by the anti-inflationary potential of anycredit contraction and accompanying downswing in economic activity. However, the return of short-term interest rates to a very low level, the supply of liquidity to markets and the weak dollar have fostered growth in prices of commodities, agricultural products and food, which have started to function to some extent as assets. Given the positive impact on inflation and inflation expectations,monetary policy is thus faced with a major dilemma.

8 The meanings of the individual currency codes in Chart II.9 can be found e.g. at www.currencysystem.com/codes/ or fx.sauder.ubc.ca/currency_table.html.

CHART II.7Monetary policy rates from the onset of the financial turbulence to the end of March 2008(%)

2

2.5

3

3.5

4

4.5

5

5.5

3/07 6 9 12 3/08

CZ EUR USSource: Datastream

CHART II.9Appreciation of the koruna against traded currencies(%, from August 2007 to March 2008)

0 5 10 15 20 25 30 35 40ILSPLNCLPCHFSKKJPYHRKBGNEURBRLDKKNOKSEKMADHUFAUDPHPSGDTNDMYRNZDCOPFJDRUBTWDCNYCADTRYKWDEGPMXNVNDIDRARSHKDGBPINRLBPUSDVEFTHBRONJMDKRWISKZAR

Source: DatastreamNote: from the start of turbulence until mid-March 2008.

CHART II.8Nominal returns on currencies for foreign investors(%; 2000–2008 average)

-505

101520

PL SK CZ HU AU ME DE CH UK US JP

USD EURSource: Eurostat, CNB calculationNote: Approximation of foreign investor's return on government bonds in the relevantcurrency. The average return is obtained as the sum of the average interest rate on government bonds and the average annual appreciation of the domestic currency's nominal exchange rate against the euro or dollar. 2008=end of March.

2 THE REAL ECONOMY

This fast appreciation, which cannot be viewed as being fully fundamental-based,also took financial market participants by surprise. Domestic and foreign analystsand forecasters are therefore forecasting a correction towards weaker values (seeChart II.10). However, there is considerable uncertainty regarding the speed andsize of the correction. Given the possible adverse effects of excessively rapidappreciation of the koruna, the CNB signed an agreement with the CzechGovernment in April 2008 on measures to prevent public sector operations fromhaving undesirable impacts on the foreign exchange market and subsequently onmacroeconomic stability.9 The trade surpluses and industrial output in January andFebruary 2008 indicated that the Czech economy and export performance wereresilient to a moderate appreciation of the Czech koruna (see Box 2 in section 2.2),although the adverse impacts of a rapid appreciation usually appear with a lag.

Except for Hungary, the macroeconomic developments in the Central Europeanregion in 2007 can be regarded as successful (see Table II.1). The macroeconomicsituation in Slovakia and Poland was favourable. Both countries should maintainrelatively high economic growth rates amid low inflation and stable externalpositions. The Central European economies are maintaining financeable currentaccount deficits, although a slight worsening is expected for Poland in 2007 and2008 (see Chart II.11). Developments in Germany, which has been showing a highdegree of resilience so far, are particularly important for the Czech economy.Leading indicators and surveys of German consumer and business sentiment aresending out by turns optimistic and pessimistic signals. Nevertheless, someweakening of the current economic boom in Germany must be expected, partlydue to rapid appreciation of the euro against the dollar and Asian currencies. TheMarch Consensus Forecast predicts that the German economy will grow by 1.7%and 1.8% respectively in the next two years, after recording 2.5% growth in 2007.

The domestic macroeconomic environment developed very favourably in 2007.Economic growth, which reached 6.6% in 2007, was positively affected by domesticdemand (increased investment activity and household consumption). The

21

9 Currency conversions of financial flows between the Czech Republic and EU authorities will continueto be effected as far as possible off the foreign exchange market. The Government will ensure that nobodies within its field of competence conduct conversions of funds inflows from the EU directly on themarket or conduct any hedging or speculative operations affecting the foreign exchange market.Furthermore, no conversions of privatisation revenues will be conducted on the foreign exchange market. If the Czech Ministry of Finance issues bonds denominated in foreign currency in the comingyears, these will be hedged against exchange rate risk.

Alternative scenario A: "Safe haven"

Scenario A, whose impacts on the financial sector are tested in section 4.2,assumes a hypothetical significant deepening of the effects of the globalfinancial market turbulence on the real economies of the Czech Republic'seuro-area trading partners. The ECB would react to the slowing economicperformance of the euro area by lowering interest rates. However, theappreciation of the Czech koruna, driven by the safe-haven motive, wouldcontinue. This would result in a decline in annual domestic GDP growth to2.5% and a slight rise in unemployment. The CNB would react by loweringmonetary policy interest rates. The decline in domestic economic growth andthe stronger koruna exchange rate would cause a rise in defaults in the non-financial corporation and household sectors, which would be moderated only partially by lower interest rates. Overall credit growth would slow radicallyand share prices would also fall, while property prices would stagnate.

CHART II.10Actual and forecasted path of the koruna exchange rate(year-on-year growth in %, outturns and September 2007 and April 2008 forecasts)

25.83 25.62

24.825.826.8

27.828.829.830.8

1/05 9 6/06 2/07 11 7/08 4/09

CZK/EUR Consensus Forecast

Source: CNB, Foreign Exchange Consensus Forecast 03/2008

CHART II.11Current account balances(% of GDP; outturns and forecasts)

-10

-5

0

5

10

15

Euro area USA Japan China CzechRep.

Hungary Poland Slovakia

2006 2007 2008f 2009f


Hungary GDP growth (%)Inflation (%)Fiscal deficit/GDP (%)

Poland GDP growth (%)Inflation (%)Fiscal deficit/GDP (%)

Slovakia GDP growth (%)Inflation (%)Fiscal deficit/GDP (%)

3.9 1.3 1.8/2.2 2.5/3.23.9 7.9 5.9/5.9 3.5/3.6

-9.3 -6.4 -4.3/4.0 -3.5/-3.2

6.2 6.5 4.9/5.3 4.5/5.01.0 2.5 4.1/4.0 3.8/3.1

-3.8 -2.8 -3.2/-3.0 -2.9/-2.8

8.5 10.4 6.6/7.4 5.6/6.34.5 2.8 3.6/3.4 3.8/3.1

-3.7 -2.6 -2.3/-2.3 -1.8/-1.8

TABLE II.1Macroeconomic indicators for Central European countries(estimates for 2008 and 2009)

2006 2007 2008 2009

Source: GDP and inflation forecasts: IMF-World Economic Outlook April 2008/EasternEurope Consensus Forecast 03/2008; fiscal deficit forecasts: OECD Economic OutlookNovember 2007/current convergence programme

contribution of net external demand was lower in 2006 than in 2007. According tothe March 2008 estimate, GDP growth in 2007 was practically the same as in 2006,not only in terms of overall growth, but also as regards growth of its individualcomponents.

The macroeconomic situation and financial sector were affected by the approvedpublic finance reform as early as in 2007 and this effect will grow in the coming years.Under this reform, a number of fiscal measures have been adopted which are havingconflicting impacts on the disposable income of households and corporations.10 Thechanges in the fiscal impulse and the impacts of the expected changes on thebehaviour of economic agents will cause some volatility in aggregate demand. A risein the VAT rate on construction work from 5% to 9% strongly affected the propertymarket and related lending. Efforts to buy real estate still at the lower VAT rate gaverise to increased demand for loans for house purchase and also affected theconstruction industry's performance. The opposite effect can be expected in thecoming years, probably compounded by a review of real estate market risks by banksin response to the problems of this market in advanced economies.

However, at the end of 2007, despite the continued robust economic growth, clearsignals of a future slowdown were observable. One of the reasons was rising inflation,which picked up markedly in late 2007 and reached 7.1% year on year in March2008.11 This is leading to slower growth in real disposable income, which is negativelyaffecting household consumption growth. The external developments and the sharpappreciation of the koruna can also be expected to dampen investment activity andnet exports in the coming quarters. The CNB's May 2008 macroeconomic forecastexpects real GDP growth to slow to 4.7% in 2008 and 4.0% in 2009. The slowingeconomic growth and declining growth of real disposable income may have someimpacts on the ability of households and corporations to repay loans taken out inprevious years. On the other hand, it may help to eliminate the excessively optimisticexpectations which emerged at the peak of the cycle and supported a rapid rise inthe indebtedness of households. It was confirmed in 2007 that the media-monitoreddata on GDP growth do not correspond exactly to the income situation of householdsand corporations, owing to a growing outflow of funds abroad. As in previous years(except 2005), gross national income growth and gross disposable income growthwere outpaced by GDP growth (see Chart II.12).

The traditional macroeconomic sustainability indicators recorded strongly positivedevelopments in 2007. The public budget deficit under ESA95 methodology fell to1.6% of GDP. As GDP growth at current prices increased by 10.1% year on year in2007, the ratio of public debt to GDP fell to 28.7%. From an international point ofview, this level is regarded as very safe. The external balance also improved (seeChart II.13). The current account deficit declined to 2.5% of GDP. The robustnessof the external position is evidenced in particular by the output balance. This turnedinto a surplus at the start of 2005 and since then the surplus has been increasingsteadily. Accordingly, its ability to cover the income deficit (which reached almost70% last year) is growing. If the CZK 120 billion in reinvested earnings werededucted from the current account balance, the current account would show asurplus of CZK 31 billion. The ratio of reinvested earnings to FDI further increasedto 70%. However, this trend is to a large extent natural and cannot be regarded asa risk from the external balance perspective.


10 A more detailed assessment of the effect of the fiscal reform on the macroeconomy was given in theCNB's October 2007 Inflation Report.

11 The high inflation in 2008 Q1 reflects an extraordinary build-up of one-off factors, namely globally rising commodity, energy and food prices and domestic increases in regulated prices, excise duties andVAT. In the course of 2008, these factors will start to unwind and inflation should return quickly tolower levels.

CHART II.12Inflation and real annual growth in economic activityindicators(%)

0

2

4

6

8

2000 2001 2002 2003 2004 2005 2006 2007

GDP GNI GDI Inflation

Source: CZSO, CNB

CHART II.13The balance of payments(% of GDP)

-8-6-4-202468

10121416

I/99 I/00 I/01 I/02 I/03 I/04 I/05 I/06 I/07

Output IncomeDirect investment Current accountFinancial account

Source: CZSO, CNBNote: Annual sliding totals of balance of payments components and nominal GDP

2 THE REAL ECONOMY

2.2 NON-FINANCIAL CORPORATIONS

Non-financial corporations posted good economic results in 2007, with many oftheir key financial indicators exceeding the three-year highs recorded in 2006.Although the sector's debt continued rising, the growth rate of loans granted tonon-financial corporations decreased over the course of the year. This may indicatea future downswing in corporate sector performance associated with the expectedweakening of economic activity due to global and domestic factors. Although the credit risk of the corporate sector remains stable, it is expected to increase in2008 in relation to the expected slowdown in economic growth. These conclusionsare indicated both by the macroeconomic credit risk model and by a newcreditworthiness indicator. Another risk to the non-financial corporations sector isthe sharp appreciation of the koruna. Although the first-quarter real economy datasuggested that corporations were able to cope with this exchange rate shock, somenegative effect on exporters can be expected to emerge gradually. The share ofproperty developers in total loans to the corporate sector has increased rapidly inrecent years against a background of rising prices of both residential andcommercial property, and currently stands at more than 25%. Given this sector'stendency to react to property price growth with a lag, a potential cooling of thereal estate market poses a sizeable risk with potential effects on the banking sector.

Profitability increased in 2007. The inventory and asset turnover ratios declined,while value added per employee and the current, acid-test and cash ratios allincreased. The debt ratio grew slightly, as did debt servicing costs owing to therising interest rates. The personnel cost-output ratio and the ratio of personnelcosts to value added rose slightly as well (see Chart II.14). Despite the increase inthe corporate debt-to-GDP ratio, this indicator for the Czech Republic is still halfthat of the EU12 countries and since 2001 has been roughly at the same level as inthe USA, where, however, corporations are traditionally financed more through thecapital market (see Chart II.15).

Despite non-financial corporations' relatively positive results, some signs of slowingcorporate sector performance are visible. These are associated with the expectedeconomic slowdown due to global and domestic factors. At the end of 2006 bankloans to non-financial corporations had shown record annual growth rates of almost21%, whereas at the end of 2007 the rate returned to a more restrained figure ofjust over 17% (see section 4.2). If the CNB's macroeconomic forecast materialises,corporate lending growth should, according to the CNB's internal models, decreasefurther in 2008, to 13%. Compared to 2006, when the growth in loans had beendriven by small and medium-sized enterprises, the growth in 2007 was veryheterogeneous. The fastest growth in bank loans was recorded for enterprises with100−249 employees. Enterprises with 250 employees or more, which can rely moreon cheaper sources of financing, such as the capital market or loans from theirparent corporations, as usual showed the lowest growth (see Chart II.16).

Despite some signs of slower growth in lending, the credit risk of the corporatesector as measured by the 12-month default rate has yet to rise and remains belowthe 3% level (see Chart II.17).12 In line with the expected economic slowdown, weassume, however, on the basis of the macroeconomic credit model, that the 12-month corporate default rate will increase by 1−2 percentage points in 2008.13

23

12 The 12-month default rate is calculated using data from the Central Register of Credits managed bythe CNB, which contains information on the bank loans of legal entities.

13 According to the estimated model, the aggregate 12-month default rate increases with appreciation ofthe domestic currency and growth in the debt ratio of the corporate sector and falls with growth in realgross domestic product and higher inflation. A detailed description of the model estimated using thedata for 1998−2006 can be found in Jakubík, P., Schmieder, C. (2008): Macroeconomic Credit riskModelling and Stress Testing. CNB Working Paper, forthcoming.

CHART II.14Key financial indicators for non-financial corporations(2006 = 100; index > 100 = improvement; index < 100 = deterioration)

7080

90

100

110

120ROE

ROA

Gross profitmargin

Debt ratio

Loans/Liabilities

Debt servicingcosts

Inventory ratio

Average collectionperiodTotal assets

turnover ratio

Current ratio

Acid-test ratio

Cash ratio

VA/Sales

Wages/VA

VA/Empl.

Interestcoverage

2006 2007Source: CZSO, CNB calculation

CHART II.15Debt ratios of non-financial corporations(% of GDP)

0102030405060708090

100

1995 1997 1999 2001 2003 2005 2007

USA EU12 Czech Republic

Source: Eurostat, US Bureau of Economic StatisticsNote: The debt ratio is calculated as the ratio of loans received and bonds issued fromfinancial accounts to GDP. The 2007 figures are estimates.

CHART II.16Credit growth(monthly data; year-on-year growth in credit to corporations by number of employees in %)

-10-505

10152025303540

10/03 4/04 10 4/05 10 4/06 10 4/07 10

1-9 employees 10-99 employees100-249 employees over 250 employeesCorporate sector

Source: CNB (CRC)

CHART II.1712-month corporate default rate by number of employees(%)

0123456789

101112

10/02 4/03 10 4/04 10 4/05 10 4/06 10 4/07 10

1-9 employees 10-99 employees100-249 employees over 250 employeesCorporate sector

Source: CNB (CRC)

The data confirm that smaller corporations are more risky. At the start of 2007, the12-month default rate was about 8% for small corporations and 3% for medium-sized corporations, while that for large corporations was less than 1%, which iswell below the total aggregate average. In addition to the corporate default rate,the financial sector is affected by the rate of recovery of claims in default, which iscorrelated with the probability of default and usually decreases as it rises. A low rateof recovery of claims has a negative impact on the financial sector. Based onMoody's data on the default and recovery rates for secured and unsecured worldbonds in 1990−2006, an inverse relationship has been estimated between thedefault rate and the rate of recovery of claims in default for secured and unsecuredbonds (see Chart II.18). Assuming that this relationship is valid for the Czecheconomy, the recovery rate for large enterprises should, based on knowledge oftheir default rate, be around 66% for secured claims and 52% for unsecuredclaims.

The hypothesis of an increase in corporate sector risk in 2007 is confirmed by thecreditworthiness indicator, which serves as an indicator of the financial stability ofthe corporate sector.14 This indicator calculates the outlook for the sector's risk atthe one-year forecast horizon based on financial indicators of solvency, profitability,liquidity and activity. The creditworthiness indicator for 2007 is somewhat lowerthan that for 2006, but is still higher than that for 2005 (see Chart II.19). Accordingto these results, corporate sector risk should show a modest increase in 2008. Thisexpectation is driven chiefly by a higher debt ratio, lower interest coverage and alower gross profit margin.

Rising prices of residential and commercial property, continued corporate investment demand, the improving income situation of households and the relatively low interest rates all helped to increase the significance of the "real estateactivities" sector (CZ-NACE 70). Strong growth in bank loans provided to thissegment of non-financial corporations in the last four years has fostered an increase in the share of property developers in total loans to the corporate sector,from 9% at the end of 2001 to more than 25% at the end of 2007 (see Chart II.20).These are mostly property development firms specialising in the construction orrenovation of commercial or residential property for renting or selling on.15

The high growth of the sector, along with the signs of a possible overheating of theproperty market (see section 3.2), are raising concerns about whether a cooling ofthe property market could generate serious problems in the sector. The highexposure of banks to developers could in turn lead to banking sector losses and theproblems could spill over via developers from the property market to othersegments of the economy (construction, etc). The default rate of developers isgenerally more volatile than the aggregate default rate and may react with a lag tomovements in property prices (see Chart II.21). The strong increase in propertyprices prior to the Czech Republic's entry to the EU may have generated optimisticexpectations in the property development sector, leading to higher investmentactivity. The subsequent sharp slowdown in property price growth may havecontributed to the later rise in the default rate in 2005−2006. The current rise inproperty prices (see section 3.2) increases the risk of a trend similar to the oneobserved around the time of EU accession. The current relatively low default rate inthis sector could thus increase again.


14 More details on this indicator can be found in the article Scoring as an Indicator of Financial Stability inthe thematic part of this Report.

15 However, they also include other real estate companies, such as estate agencies, property managementcompanies and owners' associations and housing cooperatives. The last two were not included in thedefault rate calculation.

CHART II.18Recovery rate versus default rate for unsecured worldbonds(%)

30

40

50

60

70

80

90

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5Default rate

Reco

very

rate

Source: Moody's, CNB calculation

CHART II.19Indicator of creditworthiness of non-financial corporations

0.970

0.971

0.972

0.973

0.974

2004 2005 2006 2007

Indicator of creditworthiness of non-financial corporations

Source: CZSO, CNB calculationNote: The indicator expresses the outlook one year ahead in the given year

CHART II.20Credit growth: Real estate companies versus non-financial corporations(%)

-70-60-50-40-30-20-10

010203040506070

1/01 1/02 1/03 1/04 1/05 1/06 1/07 1/08-30

-20

-10

0

10

20

30

Non-financial corporations (lhs)Real estate companies (developers, lhs)Share of loans to developers in total loansto non-financial corporations (rhs)

Source: CNB

CHART II.21Default rate of developers versus default rate of non-financial corporations as a whole and growth in flat prices(%)

0123456789

10

11/02 5/03 11 5/04 11 5/05 11 5/06 11 5/07 11-10-50510152025

Non-financial corporations (lhs) Developers (lhs)Growth in flat prices (rhs)

Source: CNB, CZSO

2 THE REAL ECONOMY

Foreign currency loans account for around 30% of total loans to developers, whichis a higher figure than for non-financial corporations as a whole. If the over-optimistic expectations regarding property price growth fail to materialise and thedomestic currency depreciates, this segment of business would be hit hard (seealternative scenario B "property market crisis").

25

Alternative scenario B: "Property market crisis"

Scenario B simulates a domestic property market crisis caused by marketsaturation. Property prices are assumed to fall by 30%, which would causeproblems in the property development sector. The large investments bydevelopers have been driven by over-optimism regarding future demand andrising property prices. If these expectations failed to materialise, loan defaultswould increase. Owing to the direct effects on other sectors (construction,etc.), as well as indirect effects, i.e. a rising unwillingness of banks to financethe real sector, domestic economic activity would decrease. Real GDP growthwould thus decline radically during 2008, unemployment would rise andinflation would fall slightly.

The CNB would react by lowering rates. The transmission to the economywould, however, be moderated by a rising risk premium on the moneymarket due to uncertainty about the amount and distribution of the losseson loans in default. The exchange rate of the koruna would depreciateslightly, further deepening the problems of developers indebted in foreigncurrency. This scenario would result in a sharp increase in the default rate,driven primarily by the corporate sector. Growth in lending would also slowand stock prices of developers and other companies would decline.

Box 2: Analysis of export-oriented corporations

The sharp appreciation of the koruna in 2007 H2 and early 2008 raises thequestion of how sensitive the Czech economy is to exchange ratefluctuations. Our analysis used financial indicators for the 611 largest Czechexporters from the Magnus database (Čekia) and data from the CNB'sCentral Register of Credits. The results show that exporters are more sensitiveto the exchange rate than is the economy as a whole. The sensitivity ofexporters varies, however, according to import use, firm size, debt ratio andthe currency structure of bank loans received.

During the 2001−2002 and 2005−2006 appreciation waves, the profitability ofthe largest exporters generally decreased faster than that of large non-financialcorporations as a whole (see Chart II.2 Box). The impact of the exchange rateon exporters' profitability is generally symmetrical, i.e. this segment is profitableat times of depreciation. Thus, greater profit volatility is a bigger problem thanexchange rate appreciation itself. In the case of the Czech economy, however,we are observing a long-term appreciation trend exerting constant pressure onexporters' profitability. The appreciation of the koruna affects the economy notonly directly via falling sales and profits of exporters, but also indirectly throughthe impact on sales of local sub-contractors and service-oriented businesses.

The appreciation of the currency may partly offset the impacts on the sales ofthose exporters who simultaneously import foreign goods for investment and

CHART II.2 (BOX)Profitability of exporters versus profitability of non-financial corporations as a whole(% and percentage points; RoE = return on equity)

-8

-3

2

7

12

2001 2002 2003 2004 2005 2006

Y-o-y change in RoE in p.p. for main exportersY-o-y change in RoE in p.p. for non-financial corporationsY-o-y percentage change in nominal effective exchange rate

CZK appreciation

Source: CZSO, Magnus, CNB calculationNote: Annual data, smoothed curves.


16 No data on specific exports and imports were available for the construction of the sample of import-hedged exporters. Information on the difference in ranking of the largest exporters in the list ofexporters and importers was used as a proxy. The set of import-hedged exporters thus included enterprises whose ranking in the list of largest importers was higher or roughly the same as that in theexporters list. For example, a firm ranked thirtieth in the exporters list and better than thirtieth in theimporters list would be included in this set.

17 The higher volatility of this indicator for exporters may also be due to the relatively small sample of corporations.

intermediate consumption. Many large exporting firms specialise in processingimported intermediate products and re-exporting them. The results of theanalysis suggest that the profitability of such enterprises is generally lesssensitive to exchange rate fluctuations (see Chart II.3 Box), although theseresults depend partly on how the individual firms are classified.16

Another factor that can affect the sensitivity of profitability to the exchangerate is company size. For large enterprises it can be easier and cheaper tohedge using currency derivatives. Foreign-controlled corporations can alsohedge via their parent companies. The results of the analysis confirm that thesensitivity of large export-oriented corporations (with 250 employees ormore) to the exchange rate is lower than that of smaller enterprises (seeChart II.4 Box).

The analysis of the currency composition of loans reveals that the exportersdrawing mostly on foreign currency (euro) loans are less risky (see Chart II.5Box). The hedging offered by foreign currency loans may thus be one of theinstruments available for reducing the sensitivity of exporting firms to theadverse effects of exchange rate movements. Exporters can also transfer partof the currency risk to their sub-contractors by means of agreements to payfor deliveries in foreign currency. The results of a CNB survey of businessessuggest that the share of foreign currency payments and receipts in totaldomestic payments and receipts has been increasing gradually since 2003. In2007, the figures were around 20% for payments and 13% for receipts.

The monthly data on repayments of loans granted to the largest exporters bybanks in the Czech Republic show that the 12-month default rate is morevolatile for exporters than for the economy as a whole (see Chart II.6 Box).17

Recently, this indicator has been lower for exporters than for the economy as a whole, but this has not always been the case. As the analysis wasconducted for relatively large corporations, values of the 12M default rateequal to or above the whole economy level indicate higher risk. However,dependence of the 12M default rate on exchange rate changes is indicatedonly in certain periods (particularly after the strong appreciation wave in2001−2002 and partly also in 2005). This would indirectly confirm thehypothesis that although the sales and profitability of exporters are sensitiveto exchange rate changes, their ability to repay loans is less dependent onsuch changes.

The debt ratio affects the sensitivity of exporters to exchange rate movements.In 2006, the average debt of the largest exporters measured as the ratio ofexternal funds to total liabilities was higher than that of larger non-financialcorporations (54% versus 45%). The results show that exporters with ahigher debt ratio (over 54%) were far more sensitive to the exchange rate(see Chart II.7 Box). This would mean that exporters' borrowings are mostly

CHART II.3 (BOX)Profitability of import-hedged exporters versus profitability of other exporters(% and percentage points; RoE = return on equity)

-4-202468

1012

2001 2002 2003 2004 2005 2006

Y-o-y change in RoE in p.p. for high-import exportersY-o-y change in RoE in p.p. for low-import exportersY-o-y percentage change in nominal effective exchange rate

CZK appreciation


CHART II.4 (BOX)Profitability of large exporters versus profitability ofmedium-sized and small exporters(% and percentage points; RoE = return on equity)

-6-4-202468

1012

2001 2002 2003 2004 2005 2006

Y-o-y change in RoE in p.p. for exporterswith up to 250 employeesY-o-y change in RoE in p.p. for exporters with over 250 employeesY-o-y percentage change in nominal effective exchange rate

CZK appreciation


CHART II.5 (BOX)12M default rate of exporters by loan currency(%; monthly data)

-6-4-202468

101214161820

11/01 11/02 11/03 11/04 11/05 11/06 11/07

Y-o-y change in nominal effective exchange rate12M default rate on CZK loans12M default rate on EUR loans

CZK appreciation

Source: CNB, CRC, CNB calculation

CHART II.6 (BOX)12M default rate of exporters versus non-financial corporations and exchange rate developments(%; monthly data)

-6-4-202468

101214161820

11/01 11/02 11/03 11/04 11/05 11/06 11/07

Y-o-y change in nominal effective exchange rate12M default rate of main exporters12M default rate of non-financial corporations as a whole

CZK appreciation

Source: CNB, CRC, CNB calculation

2 THE REAL ECONOMY

2.3 HOUSEHOLDS

Household debt continued growing, but remained lower than in the advancedWestern European EU countries. The overindebtedness of some groups of thepopulation − solving their repayment problems by taking on more debt, often athigh interest rates − poses the main risk. Owing to their low creditworthiness, thesegroups are falling into a debt trap. The non-bank sector is significant in this respect,as its loans account for about 60% of the liabilities of this group of the populationand for almost 50% of the entire consumer credit market.

Households18 create savings and provide funds to non-financial corporationsthrough banks and other financial intermediaries. Their indebtedness has increasedsignificantly in recent years and any problems with loan repayment would hit thewhole financial sector. The growing household debt has been accompanied by adecrease in the gross saving rate of households, which fell to a ten-year low in 2007(5.1%). This trend was due to an expected increase in household incomesconnected with the real convergence process, easier access to loans, the lowinterest rates in recent years, rising real estate prices and population ageing.

Household debt is increasing by roughly one-third every year. In 2007, the debt ofCzech households exceeded CZK 800 billion.19 Bank loans to households amountedto CZK 680 billion at the end of last year, of which loans for house purchaseaccounted for CZK 510 billion and consumer credit for CZK 170 billion (see section4.2). In addition, households owe almost CZK 150 billion to non-bank institutions(see Chart II.22). The share of bank loans to households in the total resident loanportfolio increased further during 2007 (to 40% in 2007 from 37% in 2006) andwas almost the same as the share of loans to non-financial corporations. This trendwas driven by households' continuing interest in buying their own homes and bythe one-off effect of advance financing of construction work owing to a change inthe lower VAT rate from 5% to 9% with effect from 1 January 2008.

The ratio of debt to gross disposable income of households was 47% at the end of2007 (compared to 40% at the end of 2006). The debt-to-financial assets ratio also

27

18 Throughout section 2.3, "household sector" refers to private individuals; in the CZSO classification, theterm households is used for both individuals and sole traders.

19 Although household debt has been steadily increasing by about one-third every year since 2002, it hasgrown significantly in absolute terms. Total household debt rose by CZK 214 billion in 2007, which is around five times the growth in 2002 and more than the total volume of household debt that year(CZK 185 billion). The absolute debt volume in 2007 was 4.5 times higher than in 2002.

in Czech koruna and not in the export currency (euro in particular). Borrowingin euro would reduce the financial costs expressed in koruna given anappreciation of the domestic currency and thus also reduce the sensitivity ofprofits to the exchange rate. The evolution of the interest rates of the twocurrencies in 2001−2006 suggests that the profitability of exporting firms wasnot driven primarily by interest expenses. An analysis of the data from thebanking credit register reveals that the share of euro loans in loans grantedto the largest exporters in the last five years is around 25% and is higher thanin the economy as a whole (about 17%). In the case of exporters, this shareis not rising much over time. The hedging function of this instrument in anenvironment of a strongly fluctuating exchange rate is probably offset bydisadvantages in the form of higher interest rates or other conditions.

CHART II.7 (BOX)Profitability of indebted exporters versus profitability ofnon-indebted exporters(% and percentage points; RoE = return on equity)

-10-8-6-4-202468

1012

2001 2002 2003 2004 2005 2006

Y-o-y change in RoE in p.p. for less-indebted exportersY-o-y change in RoE in p.p. for more-indebted exportersY-o-y percentage change in nominal effective exchange rate3M PRIBOR3M EURIBOR

CZKappreciation

Source: CZSO, Magnus, CNB, CNB calculationNote: Annual data, smoothed curves.

CHART II.22Bank and non-bank credit to households(CZK billions)

0100200300400500600

1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

Bank loans for housing purposes and non-residential real estate

Consumer and other bank credit (including overdrafts)

Non-bank credit of a consumer nature

Source: CNB, CLFA

indicates that the debt burden of households is rising (30% at the end of 2007versus 26% at the end of 2006). Total household debt as a percentage of GDPcontinued rising and is now roughly 23%. As in 2006, total interest received fromhouseholds was higher than interest paid in 2007, but the trend of slightly fallingnet aggregate interest income continued (see Chart II.23). Despite the rapid growthin Czech households' debt, the total debt level is still low compared to WesternEuropean countries, as confirmed by all three indicators mentioned above. Forexample, the ratio of debt to household disposable income exceeds 170% in theUK, 120% in the USA, 130% in Ireland and 200% in Denmark and theNetherlands. The debt-to-financial assets ratio also confirms the lower debt ofCzech households, although the differences here are not so large − Denmark 55%,Spain 47%, Portugal 47%, the Netherlands 39%, Germany 35% and Austria 35%.For some countries this indicator is in fact lower − e.g. Belgium 19%.20 Thegenerally lower debt of Czech households is also confirmed by the debt-to-GDPratio, which is about one-third of the euro area average. Hence, Czech householdsas a whole cannot be regarded as overleveraged yet. If, however, borrowingscontinue rising at the same pace, the ratio of debt to gross disposable income willreach about 57% at the end of 2008, which is the level reached in the euro areaabout ten years ago. If the debt continues growing at the same rate thereafter, thisindicator will, other trends being unchanged, reach the present euro area average(about 100%) probably at the end of 2011.21

Despite the gradually rising household debt, we are not currently observing anygrowth in the credit risk of households, which is around 3% (see Chart II.24).22

According to the CNB's internal model, we expect this indicator to stay close to thepresent level in 2008.23 Should any of the less favourable scenarios materialise, thedegree of risk could increase by about 0.5 percentage point. This applies inparticular to low-income groups. These groups may have problems with repayingtheir obligations as a result of an increase in necessary consumption expenditure in2008 bolstered by one-off effects of the public finance reform.

The existing studies analysing the equilibrium level of debt of the private sector (as measured, for example, by the ratio of loans to the private sector to GDP)confirm that the rise in debt in the Czech Republic is in line with the country'soverall economic growth (see Chart II.25). According to an Austrian central bankstudy using IMF data, this ratio could be 70% but is about 40% at present. Thismeans that there is no "excessive" growth in lending that might mask certain risks


20 All the above household sector debt indicators are calculated on the basis of aggregated balancesheets. Households may have a large quantity of financial assets on aggregate, but the same need notapply to indebted individuals. The lower values of the debt-to-financial assets ratio in some countriesmay reflect greater imbalances in wealth distribution. This effect is likely to be more pronounced for anindicator based on financial assets than for an indicator based on disposable income. In countries withseveral wealthy individuals holding a large proportion of the stocks of large firms, the debt-to-financialassets ratio may be highly biased.

21 These figures are based on relatively pessimistic scenarios regarding income and lending growth (gross disposable income growth of 7% a year and growth in loans of 30% a year). The real situation,however, may be more favourable.

22 The 12-month default rate of households can be calculated using data from the banking register ofclient information. The time series has been collected under an agreement between the CNB and theCzech Banking Credit Bureau only since August 2007. Historical data are not available in the register,so we cannot yet observe the 12-month history and the figures are estimates.

23 The macroeconomic model of household credit risk estimated for the Czech economy can be found inJakubík, P. (2007): Credit Risk and Stress Testing of the Banking Sector in the Czech Republic. FinancialStability Report 2006, pp. 61−62. This model estimates credit risk on the basis of real interest rates andunemployment in the economy, which has a direct impact on households' disposable income.

CHART II.23Ratio of debt to gross disposable income, financial assetsand GDP; ratio of interest paid to households' gross disposable income(%)

05

101520253035404550

1995 1997 1999 2001 2003 2005 2007-10

123456

Debt/gross disposable income (left-hand scale)Debt/financial assets (left-hand scale)Debt/GDP (left-hand scale)Interest paid/gross disposable income (right-hand scale)Net interest received/gross disposable income (right-hand scale)

Source: CNBNote: 2007 refers to Q3.

CHART II.2412-month default rate of households by credit type(%)

2.00

2.50

3.00

3.50

4.00

8/07 9 10 11 12

Consumer credit Loans for house purchase

Total credit

Source: Czech Credit Bureau, CNB

CHART II.25Deviation from equilibrium private sector credit-to-GDPratio(percentage points; blue curves indicate 95% confidence interval)

-80

-60

-40

-20

0

20

40

60

94 95 96 97 98 99 00 01 02 03 04 05 06

Source: Backé, P., Égert, B., Walko, Z. (2007): Credit Growth in Central and Eastern EuropeRevisited. OeNB Focus 2/07, pp. 69−77.Note: Negative values indicate that the current debt ratio is lower than the estimated equilibrium debt level measured as a percentage of GDP.

2 THE REAL ECONOMY

to the future stability of the financial system. This conclusion does not applygenerally to all Central and Eastern European countries.24

Household borrowing may have different effects on different groups of thepopulation. Low-income households and households with one economically activemember are exposed to the highest risk of potential repayment problems in thefuture. For example, according to CZSO data available for 2006, in 4% ofhouseholds with mortgages the main economically active person is unemployed (seeChart II.26). The most frequent mortgage recipient in the Czech Republic is ahousehold with two economically active persons and one child. The mainbreadwinner is a 39-year-old employee with secondary education. His partner is a33-year-old employee or housewife with secondary or basic education. About 13%of households have mortgage loans (8% of them live in rented apartments, the vastmajority of which pay regulated rents). Half of all households have net monthlymoney income of less than CZK 20,00025 while the figure for households withmortgages is CZK 27,000. The distribution of households with and withoutmortgages confirms that mortgage loans tend to be demanded by households withhigher net income (see Chart II.27). One factor of future demand for mortgage loansis the proportion of persons living in rented apartments, which was about 23% atthe end of 2006. The majority (around 80%) were in apartments with unregulatedrents. Rising interest in home ownership can thus be expected going forward.

29

24 See Backé, P., Égert, B., Walko, Z. (2007): Credit Growth in Central and Eastern Europe Revisited. Focuson European Integration 2/07, OeNB. The analysis finds that Latvia and Croatia, for example, have elevated debt levels relative to their economic fundamentals.

25 Including pensioners, who account for about 15% of households.

CHART II.26Breakdown of households with mortgages by social group(%)

31%16%

40%

1%4%

2%6%

Employees with lower educationSelf-employedEmployees with higher educationHouseholds of inactive persons with EA membersPensioners (non-working) with no EA membersUnemployedOther households with no EA members

Source: CZSO

CHART II.27Distribution of net monthly money income of households with and without mortgages(x-axis: net monthly money income in CZK thousands; y-axis: % of households ingiven category)

0

5

10

15

20

25

10 15 20 25 30 35 40 45 50 55 55+

Households with mortgages Households without mortgages

Source: CZSO

Box 3: The debt burden on households and loan repayment problems

The CZSO survey "Household income and living conditions in 2006" revealsthat the share of households using consumer credit remained at 23%,whereas that of households with loans for house purchase increasedcompared to the previous year to 11%. The percentages of low-income andhigh-income households obtaining credit were similar, while the share ofmedium-income groups was much lower (see Charts II.8 and II.9 Box). Lower-income households continued to slightly prefer consumer credit, whereashigher-income households mainly demanded loans for house purchase, butused consumer credit, which is probably partly linked with the furnishing oftheir new homes. The lower use of loans by medium-income households canbe explained primarily by the fact that they largely cover their consumptionspending from current income and do not take out loans for house purchasedue to budget constraints. Since loans for house purchase represent two-thirds of total loans in terms of volume, high-income households have alarger volume of loans than low-income and medium-income ones.

The debt burden expressed by the ratio of repayment of principal and interestto net money income was about 4% in 2006 and was roughly the sameacross all income groups of households. Repayment problems were reportedby less than 3% of households. These problems were related mainly toconsumer credit, which constituted a heavy burden for 6% of householdsand some burden for 15% (see Charts II.10 and II.11 Box). The biggestincrease was recorded in the number of households for which consumercredit repayment represented some burden. Loan repayments constituted aheavy burden for lower-income households (as reported by the two lowest-

CHART II.8 (BOX)Shares of households with consumer credit in individualincome groups(%)

05

10152025303540

1stdecile

2nddecile

3thdecile

4thdecile

5thdecile

6thdecile

7thdecile

8thdecile

9thdecile

10thdecile

2005 2006

Source: CZSO

CHART II.9 (BOX)Shares of households with loans for house purchase inindividual income groups(%)

02468

1012141618

2005 2006

1stdecile

2nddecile

3thdecile

4thdecile

5thdecile

6thdecile

7thdecile

8thdecile

9thdecile

10thdecile

Source: CZSO

One of the manifestations of the rising indebtedness is marked growth in thenumber of executions ordered in 2007. This growth also reflects the fact that thisinstrument is starting to be used to recover a wider range of claims (see Chart II.28).Most cases involved minor claims associated with more radical recovery by creditors.The overindebtedness of low-income groups of households, along with insufficientfinancial education, still poses a risk, despite the relatively stable position of thehousehold sector. Many people overestimate what they can afford and are unableto repay their debts. They resolve this situation by borrowing more, often at veryhigh rates due to their low creditworthiness, and so fall into a debt trap. The non-bank sector is significant in this respect. Loans provided by this sector account for about 60% of the liabilities of this group of people. Some 86% of all loans provided to households by non-bank institutions are for consumption (CZK 126 billion at the end of 2007). This represents almost 50% of the consumercredit market (the remainder being bank loans). This market segment is governedprimarily by the Consumer Credit Act, which now obliges creditors to informdebtors about their annual percentage rate of charge (APRC). In addition to thepoor financial literacy of debtors, this market's problems include unfair businesspractices by lenders. Further planned measures to improve consumer protectionshould help mitigate this problem. Legislative measures in the area of loanagreement conditions, e.g. the banning of arbitration clauses, would also behelpful.26 The planned implementation of financial education into primary andsecondary school curricula should also have a positive effect.

People in difficult situations leading to indebtedness can get help from several non-profit institutions and citizens advice centres. These include the SPES, theAssociation of Citizens Advice Centres, the Consumer Protection Association, theCzech Consumer Association and People in Need. According to information fromthe SPES27 and the new Financial Difficulties Advisory Centre,28 the number ofpeople seeking help from these organisations is gradually rising. The most frequent


26 These clauses, implemented by some lenders in loan agreements, mean that any future disputesbetween the lender and the debtor will be settled by arbitration. Their problematic feature is the subsequent unilateral appointment of the arbiter by the lender under the terms and conditions of theloan agreement.

27 More details about the SPES can be found at: http://www.pomocsdluhy.cz.28 The centre was opened in January 2008. More details can be found at: http://www.financnitisen.cz.

income groups − 15.8% and 11.8% respectively, although some burden wasagain found for all income groups).

According to subjective assessments, 68% of households had difficultymaking ends meet with their income, which is slightly more than in theprevious period. In the three lowest-income groups the figure was 80%−90%,whereas in the highest-income group it was only 30%. However, 41.3% ofhouseholds could not afford to pay an unexpected expense of CZK 6,000,down by 2.8 percentage points from a year earlier, but the figure for the twolowest-income groups is 60%−78%. These households are moreoverdependent on social benefits. In the two lowest-income groups, such benefitsaccounted for 18% and 7% of net money income, showing a year-on-yeardecline in the period under review. Overall, the share of households obtainingcredit increased in 2006. Loan repayment problems were reported by arelatively small percentage of households. In the two lowest-income groups,however, the figure was above the average for households as a whole.

CHART II.10 (BOX)Problems making consumer credit repayments in individual income groups(%)

0

2

4

6

8

10

1stdecile

2nddecile

3thdecile

4thdecile

5thdecile

6thdecile

7thdecile

8thdecile

9thdecile

10thdecile

2005 2006Households, total, 2005 Households, total, 2006

Source: CZSO

CHART II.11 (BOX)Problems making housing loan repayments in individualincome groups(%)

0.0

0.5

1.0

1.5

2.0

1stdecile

2nddecile

3thdecile

4thdecile

5thdecile

6thdecile

7thdecile

8thdecile

9thdecile

10thdecile

2005 2006Households, total, 2005 Households, total, 2006

Source: CZSO

CHART II.28Execution proceedings(thousands)

050

100150200250300350400450

2001 2002 2003 2004 2005 2006 20070

2

4

6

8

10

12

No. of executions ordered (left-hand scale)No. of executions completed (left-hand scale)No. of appeals (right-hand scale)Executions for non-cash payments (right-hand scale)

Source: Chamber of Executors of the Czech Republic

2 THE REAL ECONOMY

visitors to such centres are men aged 30−40,29 married with two children onaverage, with vocational or secondary education, an average monthly income ofCZK 15,000−25,000 (see Chart II.29), 12 creditors and debts of CZK 300,000−CZK 600,000 (see Chart II.30). In the vast majority of cases (95%), the clients ofsuch centres initially borrowed from banks, then moved to non-bank institutionsand, in about 50% of cases ended up turning to advertisers offering loans to lesscreditworthy clients at very high interest rates. Only about every thirtieth clientseeking help has a mortgage loan, but American mortgages and loans secured byproperty are very frequent. People ask most often about how to deal with multipleloans after misjudging what they could afford and ending up insolvent and facedwith court cases and distraint. Other questions concern the option of debtdischarge via personal bankruptcy.

The new Insolvency Act,30 which took effect in January 2008, introduces the optionof debt discharge (personal bankruptcy) for private individuals who are unable tomanage their debts. Two basic mechanisms have been defined to this end. The firstoption involves selling off the debtor's assets, and the second entails paymentaccording to a five-year payment plan. Both are conditional upon payment of atleast 30% of the amount due and on the debtor having "honest intentions" todischarge his obligations. The first case involves selling off all existing assets,whereas in the second case the debtor keeps his assets and repays the debt out offuture income. This, however, only applies to unsecured debts. Secured creditorsare satisfied preferentially through enforcement of the relevant pledges. Thedischarge method is decided on by the creditors by a simple majority of the votescast. The debtor thus cannot choose the alternative that is more advantageous tohim. Both cases end with the debt being completely cleared and at least 30% ofthe unsecured liabilities being repaid. The Act, however, allows less than 30% ofthe debt to be repaid to creditors who express their written consent. Creditors mayalso recover less than 30% if the debtor is unable − through no fault of his own,for example due to illness − to pay the agreed repayment amounts in compliancewith the five-year payment plan. In such case, the court may decide to clear thedebt at the end of the five-year period without the 30% threshold having beenmet. Personal bankruptcy creates better conditions for both debtors and creditors.Via the discharge process, it motivates debtors to repay their debts and allows themto become economically active again.31

In practice,32 personal bankruptcy is usually applied for by people with an averagemonthly income of about CZK 20,000, debts of around CZK 800,000 and 12creditors on average. The majority are divorced or single and aged either over 55 oraround 30 years. 40 years of age tends to be the exception. The applicants havemost difficulty completing the petition.33 The insolvency procedure itself incurs nodirect financial expenses, but the debtor must pay CZK 900 each month to theinsolvency trustee. The aforementioned non-profit institutions help people toprepare for personal bankruptcy.

31

29 In the case of the SPES this is just an estimate, as clients are not asked to give their date of birth. Bycontrast, the Financial Difficulties Advisory Centre does ask about the age of its clients.

30 Act No 182/2006 Coll., on Insolvency and Methods of Resolution Thereof (Insolvency Act).31 Another public source of information for potential creditors is the newly established insolvency register

administered together with the bankrupts database by the Czech Ministry of Justice. In the first threemonths after the amendment came into effect, the courts permitted 135 discharges based on datafrom this register. The experience to date, however, shows that only a fraction of people with excessivedebts meet the legal conditions for permitting the discharge process.

32 In the first three months, the SPES prepared 41 persons for personal bankruptcy, 15 of whom weregranted permission for debt discharge. Eight applications were refused and the remainder are in proceedings. Personal bankruptcy only applies to private individuals and cannot be used to dischargedebts associated with business activities.

33 The petition alone runs to 11 pages.

CHART II.29Distribution of clients of Financial Difficulties AdvisoryCentre by monthly income(% of all clients of Centre)

010203040506070

0-15 15-25 25+

Income in CZK thousands

Source: Financial Difficulties Advisory Centre

CHART II.30Distribution of average debt of clients of FinancialDifficulties Advisory Centre(% of all clients of Centre)

0

1020

30

4050

60

1-299 300-599 600-999 1000+

Average debt in CZK thousands

Source: Financial Difficulties Advisory Centre

34 Counterparty default risks on the interbank market can be measured using the spread between interbankrates and O/N swap rates at the same maturity. An O/N swap is an interest rate derivative where one counterparty pays the O/N rate on a daily basis and the other counterparty pays a fixed agreed (swap) rateon the agreed principal. In O/N swaps, the counterparties exchange only the net balance of the two rates,so the impact of counterparty default is very small and the credit risk is thus marginal, unlike in the caseof interbank deposits, where the whole principal deposited with the counterparty is at stake.

3 ASSET MARKETS AND THE FINANCIAL INFRASTRUCTURE

3.1 THE FINANCIAL MARKETS

The US subprime mortgage crisis, which started in summer 2007, had a strongeffect on global financial markets. The impact on the Czech markets was relativelylimited and largely similar to the corrections seen in the previous two years, i.e.falling stock prices, stable long-term yields and appreciation of the koruna againstthe main world currencies. Nonetheless, some indicators suggest that theincreasing global risk aversion was partially reflected in the Czech financial markets.

In mid-2007, the advanced financial markets experienced strong shocks (seesection 2.1). In July, in an environment of rising interest rates and slowing propertyprice growth, a surge in subprime mortgage defaults in the USA was reflected in a slump in the prices of bonds backed by these risky mortgages in all ratingcategories (see Chart III.1). The subsequent massive sales of risky structuredproducts led to a virtual halt in trading in these products and caused an increase inglobal risk aversion. This affected other credit derivatives markets, particularly theCDS contracts market, where spreads widened considerably (see Chart III.1). Thehigher risk aversion was instantly reflected in prices of other risky assets and inrising volatility on stock, bond and foreign exchange markets.

The preceding several years of low volatility, global excess liquidity and search foryield had encouraged financial institutions, including many European banks, toinvest in complex structured financial products, hedge funds and − via off-balancesheet exposures and credit facilities − in structured investment vehicles (SIVs,conduits) investing in bonds backed by risky assets. At the beginning of August, thesudden fall in the value of these investments raised strong concerns regarding theextent and particularly the distribution of the losses across individual banks on bothsides of the Atlantic. This resulted in an increase in the perceived potential risk of counterparty default and in the interbank rates of major world currencies abovethe level of expected monetary policy rates (see Chart III.2).34 The uncertaintyregarding bank losses was strengthened by the activation of credit lines promisedby banks to SIVs, which were no longer able to refinance themselves on the moneymarket by means of short-term bonds. The increase in interbank rates also had anegative effect on banks that had no exposure to the US mortgage market and therelated bonds and that only used the interbank market to finance their activities(e.g. Northern Rock in the UK). Since the interest rates at which the real sector isfinanced are often linked to interbank reference rates, the interest rate conditionsin the real economy tightened significantly.

The increased risk of counterparty default on interbank markets resulted in a sharpfall in liquidity on the money markets at maturities longer than about 2 weeks,requiring massive interventions by monetary authorities in order to bolster marketliquidity. The US Fed and other key central banks (the ECB and the Bank of England) reacted to the deteriorating money market conditions by increasing the liquidity offered in repo operations and later by changing the conditions for such operations, extending the maturities used to as much as 6 months and by coordinating the supply of liquidity internationally. The collateral accepted by

32 3 ASSET MARKETS AND THE FINANCIAL INFRASTRUCTURE

CHART III.1Developments on markets of subprime mortgage−backed securities and CDS contracts(CDS indices in basis points; ABX-HE 2006-2 index by rating, in % of nominal valueof backed securities)

-200-180-160-140-120-100-80-60-40-20

020406080

100120140160180200

1/06 7 1/07 7 1/080

20

40

60

80

100

120

140

160

180

200

iTraxx Europe (lha) DJ CDX IG (lha)AAA (rha) AA (rha)A (rha) BBB (rha)BBB- (rha)

Source: JP Morgan ChaseNote: The ABX-HE 2006-2 index includes bonds (incl. CDOs) backed by subprime mortgages granted in the second half of 2006.

CHART III.23M money market rates and credit premium(money market rates in %, credit premium in percentage points)

0

1

2

3

4

5

6

1/07 4 7 10 1/08 4/080.00.20.40.60.81.01.21.41.61.82.0

Euro area (3M EURIBOR, lha)US (3M LIBOR, lha)3M credit premium (EUR, rha)

1 August 2007

Source: BloombergNote: Dotted lines denote market expectations of 3M money market rates derived from FRA contracts as of 1 Aug 2007 (in case of USD from LIBOR contracts). 3M credit premium is spread between 3M EONIA swap rate and 3M EURIBOR.


central banks in these operations was also gradually extended. The Fed, and lateralso the Bank of England, introduced the option of borrowing government bondsagainst collateral in the form of mortgage-backed bonds, which can be interpretedas a measure also aimed at stimulating markets other than the money market. Inaddition, the Fed and partly also the Bank of England responded to the unfoldingcrisis and its expected impacts on the real economy by radically reducing monetarypolicy interest rates (see section 2.1), and both these banks made active use of theirposition as lender of last resort to save renowned financial institutions (NorthernRock in the United Kingdom and Bear Stearns in the USA).

Unlike the corrections in the previous two years (see Financial Stability Report2006), the global financial market situation has yet to return to normal. Betweenthe start of the crisis in summer 2007 and April 2008, the markets were verysensitive to the publication of global financial institutions' losses on risky subprimemortgages, news about the impact of the crisis on the real economy and themeasures taken by monetary and supervisory authorities to bail out probleminstitutions and calm the situation. This period saw several dramatic falls on globalstock markets, widening spreads on risky bonds, rising risk premia and volatility ona number of markets, and faster depreciation of the dollar (see Chart III.3). Long-term government bond yields in both the USA and the euro area fell in reaction toa "flight to quality" and to expectations that the financial crisis would have someimpacts on the real economy. Some segments of the credit derivatives market areessentially non-functional. Asset prices reflect the insufficient market liquidity ratherthan credit risk, which further complicates the valuation of such instruments in thebalance sheets of financial institutions and hence also the calculation of the totallosses due to the credit crisis.

The Czech financial markets responded to the developments in global financialmarkets similarly as during corrections in previous years. Share prices went down inline with the falls in foreign stock markets, while bond yields initially declinedslightly and then stabilised around 4.5% in March 2008 (see Chart III.4). Thisconfirms the past experience that Czech koruna bonds are considered a relativelysafe investment compared to the bonds of some other Central Europeaneconomies. A downward revision of the expected future monetary policy interestrate path, in an environment of a potential cooling of the economy and anappreciating domestic currency, might also have played a role.

Nevertheless, it cannot be said that the global financial crisis had no major effecton the Czech financial markets. During 2008 Q1, market signals indicatedincreased sales of Czech government bonds by foreign investors, fostering amodest increase in long-term yields. Owing to declining yields on Germangovernment bonds, the long-term spread widened gradually. The spreads on euro-denominated Czech government bonds against euro area benchmark bonds decreased slightly at the beginning of the crisis and then started growing in reaction to the rising global risk aversion, similarly to the spreads on Hungarianand Polish eurobonds (see Chart III.5). Despite this increase, Czech eurobonds are maintaining some distance from Polish and Hungarian bonds. This partlyreflected the Czech Republic's better long-term sovereign rating, which indicates its ability to repay foreign-currency liabilities (see Chart III.6).35

Given the heavy participation of foreign banks in the Czech financial sector,concerns might arise about potential spill-over of the problems caused by the losses

33

35 Rating issues are dealt with in the article The Role of Ratings in Financial Sector Stability Assessment inthe thematic part of this Report.

CHART III.3Turbulence on global financial markets: risky assets, USDexchange rate and volatility (EMBI Global, volatility indicator and USD/EUR as index, 1 Aug 2007=100; CMAXindex = ratio of current value of stock index to maximum value in past 60 days)

0.5

0.6

0.7

0.8

0.9

1.0

1/06 5 9 1/07 5 9 1/086080100120140160180200220

CMAX index MSCI Developed World (lha)CMAX index MSCI Emerging Markets (lha)EMBI Global (rha)Financial market volatility index (rha)USD/EUR (rha)

Source: JP Morgan, BloombergNote: EMBI Global Index– weighted spread of yields on dollar-denominated emerging markets bonds; Financial market volatility index – sum of historical volatility of S&P500, DJEurostoxx, 10Y US and German bonds, EUR/USD exchange rate and JPY/USD exchangerate over past 90 days.

CHART III.4The 2007 crisis: impact on stock and bond markets in theCentral European region(CMAX index = ratio of current value of the stock index to maximum value in past60 days, lha; 10Y yields in %, rha)

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

1/06 4 7 10 1/07 4 7 10 1/08 43.0

4.0

5.0

6.0

7.0

8.0

CMAX PX (CZ) CMAX BUX (HU) CMAX WIG (PL)CZ 10Ybond yield

HU 10Ybond yield

PL 10Ybond yield

Source: Bloomberg

CHART III.5The 2007 crisis: impact on 10Y yield spreads on bondsdenominated in euro(in basis points)

0

20

40

60

80

100

120

1/06 4 7 10 1/07 4 7 10 1/08

CZ HU PL

1 August 2007

Source: JP Morgan, Bloomberg

CHART III.6Development of sovereign rating of selected countries(long-term rating in foreign currency, Fitch)

1997 1999 2001 2003 2005 2007

CZPolandHungary

Germany, Netherlands, AustriaPortugalBelgium

AAA

AA

BBB+A

Source: Fitch

due to the current crisis from parent institutions to their subsidiaries. Short-termkoruna rates increased in the first phase of the turbulence in line with the expectedmonetary policy tightening. The money market recorded a modest increase in thecredit risk premium, although a much smaller one than in the euro area (see Chart III.7). Market liquidity also declined in this market (see Box 4).


36 A detailed definition of a liquid market (as applied to the stock market) can be found in Black, F. (1971):Towards a Fully Automated Exchange, Part I. Financial Analysts Journal, pp. 29-34.

37 Take, for example, a bank facing a balance-sheet liquidity shortage. The bank will try to sell assets onthe market to raise the necessary funds. If, however, it faces an illiquid market, the intended sale maybe very difficult to realise and the prices of its assets may fall under supply pressure. In the extreme case,this may trigger a spiral of rising sales of assets to raise additional funds and a related further declinein prices.

CHART III.7Czech koruna money market development(rates in %, lha; credit premium in percentage points, rha)

-1

0

1

2

3

4

5

6

1/07 4 7 10 1/08 4-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

CZ (3M PRIBOR) CZ repo rate 3M credit premium

1 August 2007

Source: Bloomberg, ReutersNote: Dotted lines denote market expectations of 3M money market rates derived fromFRA contracts as of 1 Aug 2007. 3M credit premium is spread between 3M CZEONIA swap rate and 3M PRIBOR.

Box 4: Financial market liquidity

Financial market liquidity (market liquidity) is generally regarded as the abilityof market participants to execute financial transactions in assets of a givenvolume without causing a significant change in their prices.36 The term"liquidity" can also be encountered in the context of the need to safeguardliquidity within a particular financial institution (see Box 7). It refers tobalance-sheet liquidity, i.e. an institution's ability to meet its immediatecommitments. The two concepts of liquidity are more or less interconnected.The stability of a financial institution is closely linked to financial marketstability through liquidity risk management by the institution and, conversely,through the provision of liquidity to the market via its participation in thismarket.37

This box focuses on the derivation of market liquidity indicators − ancomposite indicator for the entire Czech financial market as a whole andseparate indicators for the individual markets. It is based on calculationscontaining key information across the selected markets (the money, bond,foreign exchange and stock markets) and also across the separate attributesof market liquidity, including market tightness, depth and resilience and theliquidity risk premiums. Market tightness (short-run position turnover costs)can be measured by the width of bid-ask spread. In general, a market makeron an illiquid market will increase the width between bid prices and ask pricesas compensation for the difficulty of realising a prompt sale of an asset heholds. Market depth (the ability to execute large transactions withoutexcessively affecting the current market prices of an asset) and marketresilience (the speed at which prices recover from a random shock) can bedetermined using indicators based on the ratio of the yield to the transactionvolume. The final market liquidity calculation included here is an estimate ofthe liquidity premium. This can be understood as a form of compensationdemanded by an investor for the potential risk of having to abandon theposition associated with uncertain future market conditions. The individualsub-indicators were normalised before aggregation.

The composite market liquidity indicator (liquidity index) for the Czech financialmarket indicates relatively sharp rise in financial market liquidity between theCHART III.1 (BOX)

Composite market liquidity indicator for the Czech financialmarket

-0.5-0.4-0.3-0.2-0.1

00.10.20.30.40.5

2/00 2/01 2/02 2/03 2/04 2/05 2/06 2/07 2/08

Source: CNB, Bloomberg, Datastream


The koruna's exchange rate has followed an interesting path since summer 2007.At the start of the crisis, the koruna appreciated dramatically, probably due to theliquidation of carry trades for which the koruna was used as the financing currency,in an environment of growing global risk aversion (see Box 5). The attractiveness ofthe koruna for carry trades decreased in October 2007 owing to rising interest ratesand higher exchange rate volatility (see Chart III.8). The following period, however,saw a further strengthening of the appreciation trend, driven both by a shift inforeign investment away from the depreciating dollar and towards appreciatingcurrencies such as the Czech koruna, and by domestic exporters, who in anenvironment of faster appreciation of the koruna started hedging on a mass scaleby selling euros.

35

38 The liquidity index is constructed so that its values indicate the number of standard deviations from the historical average. It is a combination of nine individual liquidity measures, three of which involvedmeasurements based on bid-ask spreads: (1) the CZK/EUR, CZK/USD, CZK/GBP and CZK/CHF exchangerates, (2) the 12 most liquid shares in the PX index, and (3) 28 government bonds; a further four fallinto the category of liquidity risk premiums estimates: (4) the spreads between interbank deposit ratesand the monetary policy repo rate, (5) the spreads between interest rate swaps and government bondyields, (6) the historical volatility of the CZK/EUR exchange rate and (7) the Czeonia reference interestrate; and the last two comprise (8) the yield-to-trade volume ratio and (9) the yield-to-market capitalisation ratio for the 12 most liquid shares in the PX index. The individual composite indicators are then the unweighted average of the individual liquidity indicators normalised on the period2000−2/2008. The whole time series is subsequently smoothed using the Hodrick-Prescott filter.

39 ECB, Financial Stability Review, June 2007, p. 81; Bank of England, Financial Stability Report, April2007, p. 18.

40 ECB, Financial Stability Review, December 2007, p. 92; Bank of England, Financial Stability Report,October 2007, p. 10.

end of 2004 and mid-2007 (see Chart III.1 Box).38 Similarly constructed liquidityindices for the euro area and UK financial markets39 suggest a similar rise in market liquidity in recent years (since mid-2003) and a rapid fall since mid-2007.40 This indicates some global basis for the evolution of this indicatorand also confirms the increased integration of European financial markets. In the period under review, market liquidity has also decreased on the foreignexchange market, stock market (see Chart III.3 Box) and money market (seeChart III.2 Box). By contrast, the Czech government bond market is relativelyliquid and its liquidity index shows the opposite trend (see Chart III.2 Box). This may indicate increased caution on the part of investors on those marketswith falling indices (a lower number of buyers and sellers on the market, a higher liquidity risk premiums, etc.), or, conversely, a search for a potentialsafe haven on the government bond market at a time of increased marketvolatility associated with rising investor uncertainty (a flight to quality).

The market liquidity index does not include an exhaustive number of marketliquidity measures, so further refinement to include other possible measurescould have an effect on it. It is also impossible to eliminate from the primarymarket liquidity index certain temporary investor behaviour that is notnecessarily related to market liquidity but may have at least some short-termeffect on liquidity (e.g. speculation on increases in monetary policy rates). Itis also important to say that the measures included remain relevant fordefining liquidity on various markets, but only at times of relative quiet. Attimes of market tension, when the price volatility is higher, the significance ofmarket participants' behaviour, which is often over-sensitive or herd-like,increases. In such a situation, market participants assign an unhealthily higherweight to market liquidity (compared to risk and yield), so it is more difficultto determine the relevant nominal or market value of an asset at such times.

CHART III.2 (BOX)Market liquidity indicators for the money and bond markets

-1.5

-1

-0.5

0

0.5

1

2/00 2/01 2/02 2/03 2/04 2/05 2/06 2/07 2/08

Money market Bond market

Source: CNB, Bloomberg, Datastream

CHART III.3 (BOX)Market liquidity indicators for the koruna and stock markets

-1-0.8-0.6-0.4-0.2

00.20.40.6

2/00 2/01 2/02 2/03 2/04 2/05 2/06 2/07 2/08

Foreign exchange market Stock market

Source: CNB, Bloomberg

CHART III.8CZK/EUR exchange rate: the role of the attractiveness ofcarry trades and dollar development(exchange rates: 1.8.2007=100,lha; carry trade index as 3M interest rate spread ofCZK against EUR for a unit of implied volatility, rha)

85

90

95

100

105

110

115

120

1/06 7 1/07 7 1/0800.050.10.150.20.250.30.350.40.450.5

CZK/EURexchange rate

USD/EURexchange rate

Carry trade index

depreciation

Source: Bloomberg


Box 5: Carry trades and the exchange rate of the Czech koruna

Carry trades became a phenomenon in 2006 H2 and 2007 H1. Thisspeculative transaction can be described generally as an investment in a high-yielding currency financed by a loan in a low-yielding currency. The classiccase was investment in currencies such as the British pound and theAustralian or New Zealand dollar financed by loans in Japanese yen or Swissfrancs. Owing to low interest rates in the Czech Republic, the Czech korunabecame another popular currency for financing such trades at the start of2007.

In practice, financial derivatives (mainly forwards and swaps) are used morefrequently for carry trades, as they enable investors to achieve the sameexposure with lower administrative demands. Credit lines are far lessburdened and swaps spreads are narrower than those for cash transactions.It is thus cheaper for an investor to create the required position. The purchaseof higher-yielding British pounds and the sale of lower-yielding Czech korunasthrough an outright forward would be a typical example of a carry tradeusing derivatives. In the future, usually two business days before the forwardmatures, the higher-yielding currency is sold and the lower-yielding currencyis purchased, thereby finalising the transaction. This strategy allows theinvestor to make a profit from the interest rate differential between the Britishpound and the Czech koruna for the time period until maturity date of theforward.

When investing in carry trades, the investor always bears the exchange raterisk stemming from future exchange rate fluctuations of the currencies used.The investor cannot hedge against such risk by using standard financialderivatives, as he would thereby forfeit profit from the interest rate differential.A change in the exchange rate can thus force the investor to abandon aninvestment in a carry trade earlier than expected, where the exchange rate losswould be higher than the gain from the interest rate differential.

It is not easy to prove the use of the koruna as a carry trade financingcurrency. One way might be to analyse the off-balance sheet of the bankingsector. As it is financial derivatives that are mainly used for investment in carrytrades, growth in such instruments should be visible in the balance sheet of

CHART III.4 (BOX)Fixed-term CZK operations vis-a-vis non-residents inbanking sector off-balance sheet(CZK billions; upper chart assets, lower chart liabilities; figures in value of underlying asset; circled areas denote March-June period in 2006 and 2007)

1,500

2,000

2,500

3,000

3,500

4,000

-4,000

-3,500

-3,000

-2,500

-2,000

-1,500

1/06 4 7 10 4 7 101/07

Source: CNB

Alternative scenario C: "Loss of confidence"

Scenario C assumes a different impact of the current turbulence on the Czecheconomy. As in Scenario A, there would be a cooling of the global economy andthus also of the euro area economies, and domestic economic activity woulddecline sharply. Global risk aversion would rise further, reversing the previouslypositive attitude towards the Czech koruna and leading to a radical depreciation(a "loss of confidence"). Although the weaker koruna would partly help theexport-oriented corporate sector, it would on the other hand generate a largeincrease in inflation, to which the central bank would react by raising rates.

This combination of macroeconomic variables would lead to a rise in thedefault rate, generated primarily by low GDP growth and high rates. Shareprices would fall and property prices would also see a modest decline.Growth in lending would slow to a very low level.


3.2 THE PROPERTY MARKET

The property market trends identified as risky in the 2006 Financial Stability Reportcontinued into 2007. Property prices in the Czech Republic rose relatively fast in2007 despite the problems on real estate markets in many advanced economies.Besides the favourable macroeconomic situation, record demographiccharacteristics also played a role. The Czech Republic saw converging prices acrossindividual regions and individual property types. The rise in property prices led to arise in the ratio of flat prices to wages and a decline in the rent return, which fellbelow the level of mortgage lending rates. Prague seems to be the most riskyregion in terms of these indicators. It also has the highest intensity of housingconstruction. The large number of flats under construction may pose risks to theproperty development sector.

37

the banking sector at times when such positions are created or closed.Unfortunately, the data on derivatives in the off-balance sheet do not allowus to observe the direction of the trade (purchase or sale) or to distinguishthem from other derivatives transactions used for other purposes (forexample, hedging against exchange rate risk).

The aggregate data for the Czech banking sector show a sizeable increase inthe volume of derivatives in spring 2007 (see Chart III.4 Box) − of about EUR 20 billion in March−June in euro terms (a rise of 20%, compared to lessthan EUR 5 billion, or 6.5%, in the same period of 2006). This can be partlyexplained by exporters' hedging interest at a time when the koruna wasdepreciating. However, this depreciation of the koruna and the increase inthe off-balance sheet might also have been partly due to the financing ofcarry trades using financial derivatives. Furthermore, when analysing thebalance sheet of the domestic banking sector, one needs to bear in mind thatit does not include all koruna transactions, but only those where the reportingbank was a counterparty to the transaction. According to the BIS TriennialSurvey and the CNB, a large proportion (one-quarter to one-half) of korunatransactions might take place off the Czech foreign exchange market.

Another way of demonstrating the use of the koruna for carry trades is toanalyse the co-movement of the exchange rates used for such strategies, i.e.the Japanese yen and the Swiss franc in addition to the koruna. Investors'sentiment in relation to these transactions often develops regardless of thecurrency from which the position is financed. The use of the yen to financecarry trades can also be proved by the International Money Markets (IMM)data published by the Commodity Futures Trading Commission − the USexchange regulator, on the difference in the numbers of speculative futurescontracts for the purchase and sale of the relevant currency. Co-movement ofthe koruna and the yen would thus indirectly confirm the role of the korunaas a financing currency (see Chart III.5 Box).

The market data show that the exchange rates of the koruna, yen and francagainst the reference currencies were relatively strongly correlated in certainparts of the year. Given the developments on global markets and the rise ininterest rates in the Czech Republic, it is likely that the koruna-financed carrytrades were terminated at the end of summer 2007 and no longer occur toany great extent now. While the opening of carry trades probably fostereddepreciation of the Czech koruna in 2008 H1, their termination in H2contributed to its appreciation.

CHART III.5 (BOX)Exchange rates of currencies used to finance carry tradesand carry trade indicator from IMM data(currencies as indices, 13 January 2000 = 100, right-hand scale; IMM data as difference in numbers of contracts for purchase and sale of JPY futures, in thousands, left-hand scale − inverted)

-250

-200

-150

-100

-50

0

50

1001/06 4 8 11 3/07 6 10 1/08

80

85

90

95

100

105

110

Speculative position in JPY JPY/USDCHF/USD CZK/EUR

Source: Bloomberg

The relatively fast property price growth seen since 2006 H2 continued during2007. Prices of flats and building plots recorded particularly high growth (a 23%rise in supply prices year on year). Prices of family houses rose more slowly (by lessthan 5%). Prices of apartment blocks, for which only 2006 data are available,recorded an absolute fall, following previous years of growth. As in previous years,the prevailing price growth was driven mainly by rising demand for housing. Thisdemand is due, for example, to rising household income, dynamic growth in loansfor house purchase in an environment of low interest rates,41 demographic factors42

and also "forward buying" factors associated with the rise in the standard VAT rateon 1 January 2008 (see section 2.1).

Since the credit crisis in the USA can be seen in direct relation to the interaction of mortgage lending and property prices (see Box 1), one can ask whether thecurrent relatively fast growth in prices of most types of property in the CzechRepublic will lead to a similar situation. The importance of finding an answer to thisquestion is emphasised by the fact that property prices have started falling in someof the other countries where buoyant growth in prices was recorded in previousyears (see Chart III.10).43 Within the EU, the Czech Republic ranks somewherebetween the countries with relatively high price growth (Spain, Ireland, the UK,France, etc.) and those where property prices have tended to remain flat in the past(Germany and Austria). However, prices have been stagnating mainly in economiesgeographically close to the Czech Republic, which are its major trading partners.The direct impacts of the mortgage crisis are thus not likely to affect the Czechproperty market quickly. This is confirmed to some extent by the conclusions in Box 6 Identifying property market bubbles.

The apartment price growth identified above implies uncertainties associated partlyalso with the inadequate quality of the property price source data. Comparing therises in prices of flats in Prague with those in the rest of the Czech Republic forvarious sources (see Chart II.11), it is apparent that the current relatively fast growthin supply prices is not completely in line with the growth in transfer prices accordingto the CZSO, which are rising much more slowly. It is possible that transfer priceswill be revised upwards or that their low growth is due to non-market effects suchas tax optimisation,44 but the increasing difference between supply prices and finaltransaction prices can be viewed as a sign of falling market liquidity or rising risk onthe property market.

Convergence of prices across regions is also apparent. This is reflected by highergrowth in supply prices for the rest of the Czech Republic compared to Prague (seeChart III.11). Convergence of prices can also be seen across individual propertytypes classed according to degree of wear and tear. This can be identified bycomparing the growth in the supply price indices for Prague according to the IRIand the CZSO, which are constructed in slightly different ways. The IRI index


41 The direction of the causality between property price growth and growth in house purchase loansremains an issue. However, analyses suggest that loan growth to prices prevails. See also the discussionof Chart III.1.5.

42 According to preliminary CZSO data, the population grew by 93,900 people in 2007. This increase iseven higher than those recorded during the 1970s "baby boom" and only slightly lower than those inthe post-war period. Unlike in these periods, when the rise was due almost exclusively to natural population growth, the current increase is being driven mainly by record immigration (83,900 persons).However, even the natural population growth was the highest in 25 years.

43 In addition to the USA, some declines in prices can also be observed in the property markets in Ireland,the UK and Spain, although for the most part these declines are not yet included in the given timeseries.

44 The data for the CZSO's property transfer prices are sourced from prices given in property transfer taxreturns.

CHART III.10Property prices − international comparison(absolute index; 1999 Q1 = 100)

90110130150170190210230250270290

1999 2001 2003 2005 2007

AT CZ UK FR IE USA ES

Source: BIS, CZSONote: Country abbreviations given in list of abbreviations.

CHART III.11Property prices − transfer prices and supply prices(absolute index; 2004 Q1 = 100))

90100110120130140150160

2004 2005 2006 2007

Prague (CZSO transfer prices) Prague (IRI supply prices)Prague (CZSO supply prices) Rest of CZ (CZSO

supply prices)Rest of CZ (CZSO transfer prices)

Source: CZSO, IRI

CHART III.9Property prices ñ transfer prices according to tax returns(absolute index; 1999 Q1 = 100)

100120140160180200220240

1999 2001 2003 2005 2007

Family houses Apartments Apartment blocksBuilding plots Index, total

Source: CZSO, CNB calculationNote: 2007 data preliminary or calculated from supply prices


measures the price of a "standard" flat,45 while the CZSO supply price index coversall types of flats, including those with less wear and tear. The larger increase inprices according to the IRI may thus suggest convergence of prices of lower-qualityflats towards those of higher quality.

The large rise in supply prices of flats together with considerably lower increases insupply rents led to a further fall in the rent return in 2007 (see Chart III.12).46

For most large cities in the Czech Republic, the rent return is below the level oflong-term bond yields and the level of interest rates on new loans for housepurchase. This means a lower rate of return and a higher degree of risk associatedwith speculative property purchases financed by mortgage loans. The decline in therent return may potentially imply an increase in the risk of loans provided toproperty developers. Such companies currently specialise in selling flats aftercompletion and are not directly affected by a decline in the rent return. However,if their assumptions regarding future prices fail to materialise or if they fail to selltheir properties, they might be forced to let them. In this case, a decline in the rentreturn coupled with higher interest rates could mean losses for them. But no majorrise in market rents can be expected in the near future, given the ongoing rises inregulated rents.

As mentioned earlier, property prices are also closely related to household incomegrowth. If property prices rise too fast relative to household income growth,households could become overleveraged and any fall in prices would then havenegative impacts on consumption and their ability to repay their obligations. Therisk of such a development is described by the price-to-income indicator (the ratioof property prices to household income, see also Box 6). Although this indicator hasrecently been rising for most regional capitals (see Chart III.13), it is still below its end-2003 highs.47 Although wages in Prague are about 25% higher than theaverage in the Czech Republic, the price-to-income ratio identifies Prague as theriskiest region. Generally speaking, regions with lower incomes and higherunemployment have lower price-to-income ratios.

Easier access of households to loans for house purchase has undoubtedly been oneof the major factors underlying the property price increases in recent years. The linkbetween property prices and mortgage loans is clear both from the developmentof these two variables over time and from the distribution of the value of mortgageloans across individual regions. Regions with a higher average mortgage amount48

also show higher prices of flats (see Chart III.14). The correlation between the twoindicators is high (0.9). Nevertheless, in terms of interpreting the increase inproperty prices, the direction of the causality between the two indicators is alsoimportant. If the property price growth is being driven by increased demandstemming from greater availability of loans for house purchase, this could beinterpreted as convergence towards the standard features of property markets inadvanced countries. Such convergence would be relatively risk-free. However, if theproperty price growth is exerting upward pressure on mortgage loans through aneed to increase the average mortgage value, the risk of a bubble on the property

39

45 A flat with wear and tear of 40% located outside the city centre, often in a prefabricated "panel" blockof flats, i.e. flats whose prices tended to be low in the past.

46 The rent return is calculated as the one-year supply rent divided by the supply price of the flat.47 The decline in the indicator in 2004−2006 was associated with a stagnation in prices of flats amid

rising income.48 The average mortgage amount is calculated from data given in the Ministry for Regional Development's

publication Selected Data on Housing. The data give the contractual principal of mortgage loans granted, which is higher than the real amount of the loan drawn. The data are gathered from surveysof banks and are thus not fully comparable with the loan data in the CNB's statistics.

CHART III.12Rent returns(averages for period in %; comparison with yields on 10Y government bond andhouse purchase loan rates)

Prague Brno OstravaLiberec 10Y yield Loan interest

3

5

7

9

11

2000 2002 2004 2006 II/07 IV/07

Source: IRI, CNB

CHART III.13Price-to-income ratios(ratio of price of 68 m2 apartment to wage for last 4 quarters)

2

4

6

8

10

2000 2002 2004 2006

CZ Prague Ústí n/LHradec Králové Brno Ostrava

Source: CZSO, CNB calculationNote: 2007 data preliminary or calculated from supply prices

CHART III.14Average size of mortgage provided to households andtransfer prices of standard apartment (69 m2) by region(2006; CZK thousands)

0

1,000

2,000

3,000

900 1,100 1,300 1,500 1,700

Average mortgage loan principal

Apar

tmen

t pric

e

Source: MRD, CZSO, CNB calculationNote: Average for Czech Republic in red

market would be higher. In this case, the effects of a bursting of the bubble onhouseholds' balance sheets and their ability to repay their loans would also behigher. It can be derived from the data on mortgage loans that the past buoyantgrowth in mortgage loans was due more to their rising number. The contributionof growth in the average mortgage amount has so far been relatively low (see Chart III.15). This confirms the generally accepted hypothesis that the causality runsfrom house purchase loans to flat prices and that the current growth in flat pricesis relatively risk-free. On the other hand, there has been a clear rise in thecontribution of the average size of mortgage loans to their total growth. This hasgrown from 5% to 20%.

With prices rising, the fast growth in housing construction continued into 2007 (seeChart III.16). The numbers of completions, starts and flats under construction roseyear on year (completions by a significant 38%), reaching record highs for theCzech Republic. As in previous years, the housing construction was concentratedmainly in Prague and the surrounding Central Bohemian Region, which accountedfor 43.3% of housing completions. In terms of completions per 1,000 inhabitants,the intensity of housing construction in these two regions was more than 2.5 times higher than in the rest of the Czech Republic (7.6 completions per 1,000 inhabitants, compared to 3.0 in the rest of the country). The growingnumber of housing completions was also reflected in an increase in the totalnumber of flats, which according to our estimates has risen by 4.2% since 2001.49

The number of flats per 1,000 inhabitants has increased from 427 in 2001 to about440.50 Most of the properties completed were intended for sale, as reflected in anincrease in the share of owner-occupied housing from about 50% in 2001 to thecurrent more than 60% (data from the CZSO's household budget survey). Thequestion is whether the market for new housing will be satisfied in the foreseeablefuture by the increasing supply and whether the large number of flats underconstruction will pose problems for developers (see the "property market crisis"scenario). This risk may be exacerbated by the aforementioned increasing gapbetween transfer prices and supply prices, the decline in the rent return and tighterbank credit standards for loans to such companies.


49 The total number of existing flats was estimated on the basis of the number of flats determined in thelatest Census of people, houses and flats (CZSO) in 2001. The number of housing completions in2002−2007 was added and the number of flats cancelled in the same period was deducted (based on the building permit statistics). However, it should be added that this is only a rough estimate. Forexample, the number of flats increased by 292,000 in 1991−2001 (based on the census), while thenumber of housing completions was only 226,000 in the same period. According to the CZSO, aboutone-half of this disproportion was due to administrative changes in existing flats (declarations of additional flats in a single house, restitutions, etc.), while one-half was due to a physical increase in thenumber of flats, for example through reconstructions of existing non-residential buildings. To whatextent such effects are significant today remains an open question.

50 This ratio is higher than in Austria (421 flats per 1,000 inhabitants in 2004), Belgium (409), Ireland (400)and the Netherlands (422) and lower than in Germany (477), France (513) and Sweden (486). Source:Housing Statistics in the European Union 2005/2006.

51 An asset price bubble can be defined simply as an explosive and asymmetric deviation of the marketprice of an asset from its fundamental value, with the possibility of a sudden and significant correction.Asset price bubbles are often caused by psychological and behavioural factors, self-fulfilling expectations,etc. The possibility of the onset of a property price bubble in the Czech Republic was mentioned, forexample, in the 2006 Financial Stability Report.

CHART III.16Apartment construction(numbers of starts, completions and apartments under construction in given year inthousands)

Completions (left-hand scale) Starts (left-hand scale)Under construction (right-hand scale)

0

20

40

60

80

1990 1992 1994 1996 1998 2000 2002 2004 20060306090

120150180

Source: CZSO

Box 6: Identifying property market bubbles

This box mentions selected approaches to identifying unbalanceddevelopments on property markets, in particular ways of identifying"bubbles".51 Property market bubbles can have serious implications for

CHART III.15Contributions to household mortgage loan growth(year-on-year increases in %)

0

10

20

30

40

50

60

12/01 12/02 12/03 12/04 12/05 12/06 12/07

No. of loans Average loan amount

Source: CNB, MRD

3 ASSET MARKETS AND THE FINANCIAL INFRASTRUCTURE 41

52 See, for example, Helbling, T., Terrones, M. (2003): Real and Financial Effects of Bursting Asset PriceBubbles. IMF World Economic Outlook, April 2003. The effects of bursting property price bubbles areaccompanied by greater output losses and last longer on average (about 4 years) than the effects ofbursting stock market bubbles (about 1.5 years).

53 The price-to-income ratio is therefore the inverse of the "rent return" used in Chart III.12 of this Report.54 For the advanced economies in the chart both indicators are related to their long-run averages for

1990−2006, while the indicators for the Czech Republic are compared with their averages for2000−2006. The shorter period for the calculation of the averages for the Czech Republic puts somelimitations on their comparability.

macroeconomic stability and the development and soundness of the financialsector. Empirical research tells us that when property price bubbles burst, theimplications for the real economy are more serious than in the case of abursting stock market bubble.52 Property market bubbles also pose a greaterthreat to a country's financial stability if the banking sector is more exposedto such assets through investment in real estate or through mortgage-backedloans.

The "practical approach" to identifying bubbles is based on analysing simpleratios and comparing them with the long-run (average) historical value of thegiven ratio. The price-to-income and price-to-rent ratios are used mostfrequently. They denote, respectively, the ratio of flat prices to householdincome (wages) and the ratio of flat prices to market rents.53 Higher values ofboth indicators imply a higher probability of a price bubble. A comparison ofthe two indicators (see Chart III.16 Box) suggests relatively overestimatedproperty prices, for example, in Spain and Ireland (the upper right-handcorner of the chart) and underestimated property prices in Japan andGermany (the lower left-hand corner of the chart). Our estimate of thecomparable54 indicators for the Czech Republic does not suggest stronglyoverestimated property prices. However, both indicators recorded sizeableincreases last year.

The use of ratios does not guarantee correct identification of a bubble, as thehypothetical fundamental value of a property includes apart from informationabout income and rent also a number of other determinants, such as growthin house purchase loans, growth in construction output, interest rates,demographic factors (e.g. population growth) and the size of the propertymarket. The shortcomings of the "practical approach" are addressed to someextent by the "econometric approach" to bubble identification, whichcompares the market value and the estimated fundamental value of theasset. The application of the econometric approach to the data for the CzechRepublic is hindered by the short time series of key indicators and theinstability of the estimated coefficients over time, which, in turn, is related tochanges in the structural characteristics of the property market (e.g. rentderegulation). The assessment of bubbles in the Czech Republic as atransition economy is further complicated by the fact that the propertymarkets here have shown signs of underestimation in the past. It is thusdifficult to distinguish whether the current price growth is due toconvergence towards the averages of the advanced economies or whether abubble is developing.

CHART III.6 (BOX)Relationship between price-to-income and price-to-rentindices for various countries(long-run average = 100, 2006 except CZ 07)

50

100

150

200

50 100 150 200

Price/rent

Price

/inco

me

DEJPCH

ES

CANOSE

USA

FI

IT

CZ(06)KR

CZ(07)

AUUK

IEDKNLNZFR

Source: Datastream, CZSO, IRI, CNB calculationNote: Country abbreviations given in list of abbreviations.

3.3 THE FINANCIAL INFRASTRUCTURE

In 2007, financial stability continued to be supported by the smooth operation of the interbank payment system CERTIS and the short-term bond system SKD,55

both of which are administered by the Czech National Bank. These systemsrecorded no irregular situations in the period of turbulence in global markets. TheCNB aims to safeguard the quality and security of the services provided andmonitors current developments and trends in the European financial infrastructure.The further development of the capital market in the Czech Republic and itsexternal competitiveness would benefit from the existence of a central securitiesdepository, but this still remains in the preparatory stage.

The Short-Term Bond System (SKD) is used for issuing and registering all book-entrysecurities with maturities of up to one year and for settling trades in thesesecurities. T-bills and CNB bills are registered in SKD. The system enables sales ofsecurities, repos and sell and buy operations, as well as pledges and exchanges ofsecurities. CERTIS (Czech Express Real Time Interbank Gross Settlement System)processes all domestic interbank transfers in Czech koruna in real time.

The volume of transactions processed in SKD was gradually rising between 2000and 2006, and amounted to almost CZK 47,000 billion in 2007 (see Table III.1). An average of CZK 185 billion was processed every day. SKD's turnover in roughly19 days equalled annual nominal GDP.

Smooth and stable interbank settlement is supported by the use of intraday credit(see Table III.1). In 2007, the volume of intraday credit grew by around 4%compared to the previous year to CZK 7,152 billion. This meant a continuation ofthe previous years' steady upward trend in the volume of intraday credit as banksbecame more aware of how it can be used. Through SKD, the CNB provides CERTISparticipants with interest-free intraday credit to boost their balance-sheet liquidityduring the day. All intraday credit extended to commercial banks by the CNB iscollateralised. The credit is used regularly by about 15 banks, five of which use it to a greater extent. While some banks use the credit almost every business day inthe month, others use it either not at all or only rarely. Therefore, the averagenumber of days on which banks used the credit was around 8 days56 in 2007 (seeChart III.17). However, it is not possible to determine whether banks used the creditmore frequently in any particular period of 2007. As regards the volume of credit,higher − but not extreme − amounts were used between June and August (seeChart III.18). It appears, then, that Czech banks did not need to use this instrumentmore often, despite the liquidity problems in global financial markets.

CERTIS ran smoothly, with a continued upward trend in the number of paymentssettled (see Table III.2). CNB Clearing processed 411 million items totalling CZK 174,854 billion, up by 15% compared to 2006. The average daily number of items processed was 1.64 million and the average daily value of the items wasCZK 697 billion (see Charts III.19−III.21). These figures reveal the extent of paymentsettlement in CERTIS and its significance for financial stability. It took roughly fivedays to reach a turnover equal to annual nominal GDP.


55 A more detailed description of these systems was published in the 2004 Financial Stability Report. Riskmanagement and the evaluation of these systems against international standards was dealt with in the2005 Financial Stability Report.

56 Average for banks that used the credit at least one day in the month.

20002001 200220032004200520062007

23,258 27,350 n.a. 22,865 22,334 n.a. 32,418 16,615 n.a.39,040 17,029 2,49340,713 16,214 3,05538,742 14,552 3,55747,534 13,810 6,88446,902 12,870 7,152

TABLE III.1SKD − statistical information

Total value oftransactions

(CZK billions)Period

Total number

of transactions

Total volume ofintraday credit(CZK billions)

Source: CNB

CHART III.18Total and average value of intraday credit(2007)

300,000400,000500,000600,000700,000800,000

4,000

5,000

6,000

7,000

8,000

Total value of IC (left-hand scale)Average daily value of IC (right-hand scale)

1 2 3 4 5 6 7 8 9 10 11 12

Source: CNB

200220032004200520062007

100,343 431 262 1.12 5.696,938 385 317 1.26 6.6

110,127 434 333 1.32 6.4123,354 488 356 1.40 6.0151,537 604 382 1.52 5.3174,854 697 411 1.64 5.1

TABLE III.2CERTIS interbank payment system − statistical information

Turnover(CZK

billions)Period

Average daily

turnover (CZK billions)

No. of

transactions (millions)

Averagedaily no. of

transactions(millions)

GDP/Averagedaily

turnover

Source: CNB

CHART III.17Average and maximum number of days in month whenbanks used intraday credit(2007)

01 2 3 4 5 6 7 8 9 10 11 12

510152025

Maximum number of days on which bank used IC

Average number of days on which bank used IC

Source: CNB


As regards number of items, non-priority items transferring lower amounts aredominant in the long term. Items with amounts of up to CZK 100,000 account for96% of all transactions in the month. By contrast, in terms of the volume of thefunds transferred, the remaining 4% of the number of items represent more than98% of total turnovers. Items with amounts of over CZK 10 million account for0.1% of the number of items and 94% of turnovers. An analysis of the monthlyturnovers of priority items in 2005−2007 reveals that ten participants in the systemare regularly the biggest payers into the system and simultaneously the biggestpayment recipients (about 87% of the total turnover). Each of them receives andsends payments to more than ten participants. The maximum share of monthlyturnover of a single payer in the total monthly turnover is 20%. The shares of thesebiggest payers are roughly the same for the comparable periods of 2005−2007 (seeChart III.22).

In order to test the contingency plan, CERTIS items were processed at the backupfacility on two days of 2007. The test confirmed that the CERTIS system can be runfully at this facility if necessary.

Central depository

The basic infrastructure of the Czech capital market for the registration andsettlement of investment instrument transactions is currently highly fragmentedand thus inefficient. One of the main reasons is that there is still no single centralentity − a central depository − maintaining a securities register and performingtransactions settlement as is the case in the advanced capital markets.

Transactions in investment instruments are settled by Univyc, a.s. (exchange andoff-exchange transactions), RM-S, a.s. (transactions concluded on the over-the-counter market it organises) and the CNB (the short-term bond market). However,all domestic book-entry and immobilised securities are still registered by the CzechSecurities Centre. Univyc maintains a register of physical and foreign securities,while the CNB maintains a register of short-term bonds in SKD. Cash settlement isperformed by Univyc, a.s. through the CNB's clearing centre and by RM-S throughČSOB, a.s. Transaction settlement is thus fragmented and associated with hightransaction costs. After the central depository opens for business, registration ofinvestment instruments and settlement of transactions, along with other relatedactivities, will be performed by this single entity, which should lead to more efficientsettlement of transactions in investment instruments, lower transaction costs andalso reduced operational and credit risk. Only entities subject to state supervisionwill be granted direct access to the central depository. This will increase the securityof the investment instrument register maintained at the central depository, and willalso reduce costs. These costs are currently high, partly because the register at theCzech Securities Centre is accessed through counters. Although discussions aboutestablishing a central depository have been going on for years, the groundwork hasstill not fully been laid for it.

Monitoring of European trends

CNB representatives work in the relevant ECB committees and working groups and monitor the development of the four key European infrastructure projects,which advanced fairly significantly in 2007: progress towards the single europayment area (SEPA), the launch of TARGET2 and CCBM2 (a single shared platformfor collateral management) and the preparation of TARGET2-Securities (a commontechnical platform for the settlement of securities transactions in central bankmoney).

43

CHART III.20Number of transactions processed by CERTIS in 2007(millions)

0.0

0.5

1.0

1.5

2.0

2.5

1/07 3 5 7 9 1128

30

32

34

36

38

Average daily no. of transactions (left-hand scale)Total no. of transactions in month (right-hand scale)

Source: CNB

CHART III.21Average daily turnover in CERTIS in 2007(CZK billions)

0

200

400

600

800

1,000

1/07 3 5 7 9 11

Source: CNB

CHART III.22Share of largest payers in total CERTIS turnover in priority items(%, ten largest payers, data for May of respective year)

0

5

10

15

20

25

1 2 3 4 5 6 7 8 9 10

5/05 5/06 5/07Source: CNB

CHART III.19CERTIS interbank payment system(number of transactions processed in 2003−2007)

0100200300400500600700800900

1/03 1/04 1/05 1/06 1/07

CZK

billio

ns

0510152025303540

milli

ons

Average daily turnover (left-hand scale)Total no. of transactions in month (right-hand scale)

Source: CNB

The aim of SEPA is to unify payments in the EU so that consumers can pay in eurounder the same basic conditions regardless of whether they are making a nationalor cross-border payment. SEPA is based on the application of uniform Europeanstandards for credit transfers, direct debits and card payments (SEPA rulebooks).The system of SEPA rulebooks for credit transfers was successfully launched inJanuary 2008. More than 4,000 European institutions joined it from the verybeginning, including several banks in the Czech Republic. The application of SEPArulebooks for direct debits is expected in 2009 and for card payments between2008 and 2010. The ultimate goal is for the critical mass of payment transactionsin the EU to migrate to SEPA by the end of 2010.

The approval of the Directive on Payment Services in October 2007, which issupposed be transposed into the law of the EU Member States by 1 November2009, is considered the key step towards the implementation of SEPA. Amongother things, the directive removes the previous uncertainty regarding the directdebit rules. The approved scheme is based on the principle that the direct debitpayment process can be only initiated by the payee, which will mean a change incurrent Czech practices.57

The Eurosystem's new payment system for large and urgent payments TARGET2was successfully launched on 19 November 2007. It is based on a single technicalplatform shared by all users. It will replace the decentralised TARGET system, whichhas been in operation since January 1999. A total of 259 banks from 8 countriesjoined the new system in the first stage. Another 13 EU countries joined in twosubsequent migration waves on 18 February and 19 May 2008.


57 In most Western European countries, a contract between the payer and payee must be submitted tothe payee's bank. The bank maintains a database of such contracts and checks payees' identificationnumbers. A different system applies in the Czech Republic. A database of mandates is administered bythe payer's bank. The bank decides whether to approve the debit from the payer's account based onwhether the payee's request matches the data in the database. This means the principle in the CzechRepublic will be reversed.

4 THE FINANCIAL SECTOR


While global financial institutions were significantly affected by the credit crisis, theCzech financial system remained fairly isolated from the global turbulence. Majorinternational banking groups were forced to admit large losses related directly orindirectly to a decline in prices of risky assets, especially bonds backed by defaultingUS mortgages.58 Czech financial institutions held a minimum amount of such riskyassets, mainly because of the strong focus of banks and other financial institutionsin the Czech Republic on the traditional (conservative) business model on the as yetunsaturated Czech market.59 This focus is reinforced by the prevailing foreignownership of domestic financial institutions, as foreign owners let their subsidiariesin new EU Member States generate income mainly from dynamically developingretail banking, while administration of securities and derivatives portfolios istypically concentrated in parent institutions or branches in financial centres (Londonand New York). The stability of the domestic banking sector in times of financialmarket turbulence has also been fostered by banks' high balance-sheet liquidity,the prevailing financing of credit expansion with primary deposits and thusminimum dependence on funds from foreign markets or parent companies.Moreover, domestic financial institutions do not belong to the global financialgroups that have been hardest hit by the crisis.

Despite the above factors, the Czech financial sector will not necessarily remainimmune to the current credit crisis. Large domestic banks with a strong surplus ofbalance-sheet liquidity might become a potential source of liquidity for their foreignparent institutions, which could subsequently affect the financing of the Czecheconomy. However, lending to parent banks is limited by the regulations. Certainmedium-sized banks that do not have a broad deposit base might limit creditexpansion and reduce the level of competition in the sector. Increased risk aversioncould furthermore cause a decline in some other riskier assets held by Czechfinancial institutions in their portfolios. Moreover, preliminary signals confirm thatsome subsidiaries of foreign banks might have tightened their credit standards dueto tighter credit policy within the globally operating group in response to the crisis.

The available analyses indicate that the Czech financial sector (and in particular thebanking sector) is not exposed to the risk of a crisis similar to the one that hit theUS subprime mortgage segment. This is due to very conservative loan-to-valueratios, traditionally higher required debtor creditworthiness, the traditional methodof interest rate fixation, less use of external mortgage underwriters and the absenceof significant credit securitisation. Nevertheless, it is vital to constantly monitor thisarea and assess any signs of increasing risks in a timely manner.

45

58 In its Global Financial Stability Report (April 2008), the IMF estimates the total losses of financial institutions related to the US mortgage market crisis at USD 945 billion. This sum is a combination of losses from loans and losses from related securities, with USD 565 billion pertaining to residentialmortgage loans and the remaining USD 380 billion to other credit market segments. In a report for the OECD Committee on Financial Markets entitled The Subprime Crisis: Size, Deleveraging and Some Policy Options (April 2008), OECD experts estimate the total losses to be considerably lower (USD 422 billion), mainly because they focus only on losses related to residential mortgages. The differences in the loss estimates are also connected with difficult-to-determine assumptions about therecoverability of non-performing assets.

59 According to a CNB survey conducted at the start of the crisis in summer 2007, the Czech banking sector's overall exposure to CDOs was about CZK 11 billion, i.e. just 0.3% of assets (of which, moreover, only about 5% were directly related to subprime US mortgages). In the case of insurance corporations and pension funds, the figures were about 0.15% and 0.5% of assets respectively.

4.1 FINANCIAL SECTOR DEVELOPMENTS

The Czech financial sector experienced mostly positive developments in 2007. Thebanking sector achieved record profits in 2007 and showed continuing stronggrowth in lending to the real economy. Insurance companies have significant long-term potential for further development in both the life insurance and non-lifeinsurance segments and hold capital above the required solvency margin. Insurancecompanies and pension funds saw marked growth in the costs of intermediatingnew contracts. These costs may negatively affect future profitability. Mutual fundsare a popular investment opportunity for households. There was growing interest infunds with distributed risks, such as mixed funds, funds of funds and foreignguaranteed funds.

The depth of financial intermediation in the Czech Republic,60 as measured by theratio of financial sector assets to GDP, increased from 133% in 2006 to 142% in2007. Financial intermediation, as measured by the volume of assets of financialinstitutions, grew by 17.5% year on year in 2007 (compared to only 7% in 2006).Within the structure of financial system assets (see Chart IV.1), there was a sharp(18%) year-on-year increase in bank assets. A similar rate of growth (19%) wasrecorded by the activities of investment companies and the domestic mutual fundsthey administer. Pension funds increased their assets by 14%. Unlike the othersectors, insurance companies recorded a smaller rise in assets (about 6.5%) due tolower growth in non-life insurance. Overall, both banks and insurance companiesachieved high returns on assets (1.3%, 3.7%) and equity (24.5%, 21.7%) in thefavourable economic growth conditions. The quality of the domestic financialmarket has improved further since the sales of the state-owned stakes in largebanks in 2001−2007, and the activities of other non-bank institutions haveundergone further development (see Chart IV.2).

In 2007, the financial sector was shaped mainly by the favourable phase of thebusiness cycle and the expected impacts of the fiscal reform. Also significant wasthe transition to the Basel II framework in the credit institution and investment firmsectors and the preparation for the new Solvency II framework in the insurancesector. The credit crisis on foreign markets in the second half of 2007 and the firstfew months of 2008 also had some impact.

4.1.1 The banking sector

2007 was another successful period for the banking sector. Total assets saw furthergrowth, fostered by a sizeable increase in loans reflecting the favourable economicgrowth and rising demand for owner-occupied housing. However, growth in housepurchase loans to households and a further rise in the rate of growth of loans toreal estate companies may become a risk element in the future (see section 3.2).The quality of loan repayment is affected by the disposable income of debtors,property prices and the level of interest rates. A negative trend in any of these areaswould probably result in a deterioration in loan quality. Banks as a whole againrecorded strong profits. Provided that a sufficient proportion of the profit remainsin banks in the form of equity capital, this lays the groundwork for maintaining thesector's stability in the future.

The intensive preparations for the implementation of Basel II and the actualchangeover to the new prudential rules in several banks on 1 July 2007 were asignificant challenge for the banking sector in 2007. The remainder of the sectortook this step in January 2008. Owing to the gradual changeover to Basel II, therewas a slight increase in capital adequacy. This reflects the fact that banks made use

46 4 THE FINANCIAL SECTOR

60 The regular Analyses of the Czech Republic's Current Economic Alignment with the Euro Area (CNB2007) contain a comparison with the monitored euro area countries.

CHART IV.1Shares in financial sector assets(%; 2007)

74.2%

0.2%7.0%

3.3%

3.8%5.5%

6.0%

BanksCredit unionsInsurance companiesPension fundsInvestment companies, investment and mutual fundsLeasing companiesOther non-bank financial corporations

Source: CNB, CZSO

CHART IV.2Growth (fall) in shares in financial sector assets2001−2007 (percentage points)

Pension funds

Insurance companies

Banks

-4 -3 -2 -1 0 1 2 3 4 5

Credit unions

Other non-bankfinancial corporations

Investment companies,investment and mutual funds

Leasing companies

Source: CNB, CZSO


of the possibility of more accurately assessing the risks they undertake. As expected,this led to a decline in capital charges61 and more efficient use of capital.

Loans and credit riskClient loans extended by the banking sector totalled CZK 1,786 billion at the endof 2007. The annual growth rate of loans reached 26.4% (see Chart IV.3), thehighest level recorded for the entire period under review, i.e. since 1996. Comparedto the end of 2006, the growth rate was up by 6.5 percentage points. However, theannual rate of growth of total loans declined by almost 2 percentage points in 2008 Q1, to 24.5%. Although the current growth remains fairly robust, it is notexcessive according to available studies (see section 2.3). As regards the creditstructure, non-financial corporations are still the main debtor of banks, with a 42%share of total loans. Loans to households accounted for 37.5%.62

Client deposits, which were 1.3 times higher than client loans at the end of 2007,are the biggest source of loan financing. Financing of client loans by primarydeposits in the Czech Republic is more than two times the average in the original EUMember States and 40 percentage points higher than the average in the new EUMember States (see Chart IV.4). At the end of 2006, the new Member States whichjoined the EU in 2004 or later held client deposits which were 10% higher than thevolume of loans granted. By contrast, the old EU Member States sought on average20% of missing funding sources for loans outside client deposits (on the interbankand capital market). It is this difference in the financing of bank assets that nowmakes the banking sectors in the new economies, including the Czech Republic,more resilient to the consequences of the mortgage crisis in the USA in the area ofbalance-sheet liquidity. The Czech Republic is one of the countries where banks'primary funds base is currently large. There are thus two advantages of this largevolume of client deposits: protection against any rapid drying up of liquidity on thefinancial market, and the low costs of such funds compared to other forms ofexternal financing. However, deposit growth (17% in 2007) has been lower thancredit growth in the Czech Republic. The ratio of deposits to loans is graduallydeclining and banks may thus lose the above advantages.

The rise in lending also means a rise in credit risk exposure.63 The ratio of defaultloans64 to total loans was 2.7% at the end of 2007 (see Chart IV.5), down by 0.9percentage point from a year earlier. This ratio decreased in all sectors of theeconomy. This was due mainly to the favourable economic environment. The highrate of growth of loans is probably currently resulting in a slight overvaluation oftheir quality, as expressed by the percentage of default loans.65

47

61 Basel II introduces a new category of capital charges for operational risk. Despite the creation of this newcategory, the banks that adopted the new rules in 2007 recorded capital savings overall, owing to generally lower capital charges reflecting more accurate measurement of other risk types. The share ofoperational risk capital charges in total capital charges is dealt with in the article Operational Risk and itsImpacts on Financial Stability in the thematic part of this Report.

62 The remaining 20.5% are loans extended by banks to the government sector, non-banking financialinstitutions, small businesses and non-residents.

63 In its dominant form, credit risk is the risk of default on a loan or part thereof, or of default on contract leading to delayed repayments. This risk is usually a subject of ratings by external institutions.This issue is dealt with in relation to financial institutions in the article The Role of Ratings in FinancialSector Stability Assessment in the thematic part of this Report.

64 A default loan is defined by CNB Decree No. 123/2007 Coll., on prudential rules for banks, credit unionsand investment firms, as exposure to a debtor in default. A debtor is in default at the moment when it isprobable that he will not repay his obligations in a proper and timely manner, without the creditor proceeding to satisfaction of the claim from the collateral, or when at least one repayment (the amountof which is deemed by the creditor to be significant) is more than 90 days past due. The term default loanis essentially equivalent to the former term non-performing loan, which was used in last year's Report.

65 The risk of loan default, as expressed by the default rate in the non-financial corporations sector, hasbeen roughly the same since 2002 (see section 2.1). The default rate for total loans to households,monitored since 2007 H2, has also been almost stable (see section 2.2).

CHART IV.3Year-on-year credit growth by sector(%)

-40-30-20-10

01020304050

2000 2001 2002 2003 2004 2005 2006 2007

Total Non-financial corporationsFinancial institutions excluding banks TradesHouseholds

Source: CNB

CHART IV.4Financing of loans by primary deposits (%)

04080

120160200

CzechRepublic

Hungary Poland Slovakia EU EU untilMay2004

NewMemberStatesafter

May 2004

Deposits/loans 2004 Deposits/loans 2005 Deposits/loans 2006

Source: ECB

CHART IV.5Default loans by economic sector (% of total loans of given sector)

0

5

10

15

20

25

2003 2004 2005 2006 2007

Total Non-financial corporationsFinancial institutionsexcluding banks

Government

TradesHouseholds

Source: CNB

Loans to non-financial corporationsDemand for products and services of non-financial corporations increases whentimes are good. Such corporations need considerable external sources of financing.The volumes of bank loans to corporations have been growing since 2004 andamounted to CZK 744 billion at the end of 2007. Manufacturing was the mostimportant debtor in terms of sectors (26%), followed closely by real estatecompanies (25%).

The rate of growth of bank loans to non-financial corporations decreased slightlyin 2007, by 3.7 percentage points to 17%. Based on the developments in the firstfew months of 2008, it cannot be ruled out that the slowdown will continue. Thebiggest contributor to the growth in 2007 was the real estate sector with 46% (seeChart IV.6), followed by manufacturing. Loans provided to real estate companieshave been growing faster and faster since 2003 H2 and showed record growth of50% in 2007. Such high growth is consistent with the large increase in demand fornew housing construction, which is being supported by a wide supply of bankloans. Given the enormous indebtedness of property developers, this sector and theproperty market were included in the "property market crisis" scenario of the stresstesting of the Czech banking sector (see section 4.2). The results of the test showthat the banking sector is sufficiently resilient to the risks associated with potentialnegative developments in developers' business activities and on the propertymarket.

The favourable economic environment is having a positive effect not only on creditgrowth, but also on the ability of the corporate sector to repay its obligations. Thequality of loans to corporations as a whole, as expressed by the ratio of defaultloans to total loans, is steadily improving. This ratio declined by 0.5 percentagepoint last year, to 3.1% in December (see Chart IV.7). Industries with the largestdebts, i.e. manufacturing and real estate companies, recorded ratios of defaultloans to total loans of 4.2% and 1.3% respectively at the end of 2007. In bothcases, the ratio of default loans to total loans declined year on year, largely due tolarge volumes of new loans. The effect of the current strong growth in loans is thatthe existing loans are relatively "young". If the rate of growth declined as a resultof an overall slowdown in the economy or saturation of the given market (a declinein demand for residential and commercial property), the average "age" of the loanswould gradually lengthen and the likelihood of repayment difficulties wouldincrease. The average "age" of a mortgage loan is about 3.5 years (see Chart IV.8).This will lengthen if the slowdown in the provision of new loans of this kindobserved in 2008 Q1 continues. Mortgage loans are now granted with a maturityof about 20 years on average.

Loans to householdsLoans to households grew by 35.1% in 2007, reaching CZK 669 billion at the endof the year. After having slowed slightly in 2006, their rate of growth picked upagain, by 4.7 percentage point year on year (see Chart IV.9). In 2007, the growthin total loans was driven mainly by house purchase loans, which accounted for76.4% of total loans to households and increased by 37.6% in 2007, i.e. 5percentage points faster than in 2006. By contrast, the rate of growth of consumercredit (26.1%) recorded a slight decline and ultimately led to a decline in the shareof consumer credit in the structure of loans by purpose to 18.8%.66


66 The CNB's internal estimates based on economic models indicate that the current increase in consumercredit is relatively sound, i.e. driven by fundamentals. According to the model, a decline in the rate ofgrowth of bank consumer credit to just above 20% can be expected in 2008.

CHART IV.6Shares of industries in total annual increase/decrease incredit to corporations(%)

-150-100-50

050

100150200

2002 2003 2004 2005 2006 2007

Manufacturing Public servicesConstruction Trade, hotels and restaurantsTransport, communications Real estateOthers

Source: CNB

CHART IV.7Default loans by industry(% of total loans of given industry)

05

10

152025

2002 2003 2004 2005 2006 2007

Manufacturing Public servicesConstruction Trade, hotels and restaurantsTransport, communications Real estate

Source: CNB

CHART IV.8Average age of a mortgage loan(years; y-axis)

1.0

1.5

2.0

2.5

3.0

3.5

2002 2003 2004 2005 2006 2007

Source: CNB

CHART IV.9Year-on-year growth in credit to households by purpose(%)

152025303540455055

2002 2003 2004 2005 2006 2007

TotalLoans for house purchaseConsumer credit (excluding current account overdrafts)

Source: CNB


New loans in 2008 Q1 signal the possible start of a slowdown in new borrowingby households. While loans to households recorded an increase of 24% in the firsttwo months of 2007 compared to the same period a year earlier, the rise in newloans to households was "just" 13% in January and February 2008 compared toearly 2007. The volumes of new loans for house purchase rose by 16.3% in thesame period (compared to 31.5% a year earlier). New consumer credit evenrecorded a slight decline in January and February 2008 compared to 2007.

In 2007, the rising indebtedness of households in the Czech Republic was fosteredmainly by growing income, ever expanding supply from banks and developers, and the persisting low interest rate environment. Factors common to the new EUMember States include a low initial level of household debt, a preference forowner-occupied housing and a visible change in households' behaviour with regardto consumer credit, as they are ceasing to be afraid of financing their short-termneeds with credit if they are short of money. The rate of growth remained higherin the new Member States than in the old Member States in 2007 (see Chart IV.10).

The ratio of default loans to total loans to households recorded an annual declineof 0.1 percentage point to 2.7% at the end of 2007. The quality of loans tohouseholds was affected by the large volume of new loans and the dominant shareof less risky house purchase loans. Loans for house purchase are the highest-qualitycomponent of the credit portfolio, as expressed by both the ratio of loans in defaultto total loans and by the default rate (see section 3.2).

Loans for house purchase ended 2007 with a 1.5% share in default loans (seeChart IV.11), which is comparable with the end of the previous year. Mortgageloans (loans fully secured by property) account for about 65% of loans for housepurchase. The loan-to-value indicator reached 56% for mortgage loans tohouseholds at the end of 2007 (compared to 53% at the end of 2006). Lower-volume loans usually take the form of building society loans not secured byproperty and special-purpose consumer credit. For several years, the share ofdefault loans has been about 1 percentage point lower for mortgage loans than forunsecured loans. It stood at 1.2% at the end of 2007.

The share of default loans in total consumer credit was 6.6% at the end of 2007,down by 0.7 percentage point year on year. The ratio of credit card credit toconsumer credit increased slightly to 8.7% year on year. Its quality is still higherthan that of total consumer credit (see Chart IV.12).

The rate fixation period is a potential risk for both banks and debtors. Variable andshort-term fixed rates responding flexibly to changes in market conditions areadvantageous for clients at a time of declining interest rates. At a time of anunexpected increase in interest rates they generate greater income for banks, whileimplying an increased burden for clients, which might, in the extreme case, resultin default. The moment of change in the rate is usually associated with the optionof early repayment, which poses a risk for the bank. If the client refinances the loanafter a short period of time at a rival bank, the transaction may, in some cases,become unprofitable for the original bank due to considerable initial costs. Thestructure of new consumer credit by initial rate fixation period (see Table IV.1) isconsistent with the structure by maturity. But the situation is different with loansfor house purchase, as the client may usually choose a short initial rate fixationperiod even for a long-term loan. At the turn of the year, uncertainty regarding thefuture evolution of interest rates caused new clients to start more stronglypreferring a long rate fixation period for mortgage loans. Building societies, whoseloan rates are fixed for the entire repayment period, are more active on the long-term loan market.

49

CHART IV.10Contributions to annual rate of growth of credit to households in EU Member States(%; 2007)

LTPL

CZSK HU

MT IE DK FR IT AT NL0

20

40

60

80

RO BG LV ET SI GR EE FI PT LU SE BE

Loans for house purchase Consumer credit Others

Source: ECB

CHART IV.11Default loans to households for house purchase(% of total loans of given type)

0.5

1.0

1.5

2.0

2.5

3.0

2001 2002 2003 2004 2005 2006 2007

Loans fully secured by propertyLoans not secured by propertyTotal loans for house purchase

Source: CNB

CHART IV.12Default consumer credit to households(% of total loans of given type)

0

24

68

10

2001 2002 2003 2004 2005 2006 2007

Consumer credit Credit card credit

Source: CNB

20062007Jan−Feb 2008

20062007Jan−Feb 2008

100 32.3 27.2 40.5 n.a. n.a.100 29.9 25.9 44.1 n.a. n.a.100 40.2 26.6 33.2 n.a. n.a.

Loans for house purchase

100 38.7 29.1 n.a. 7.8 24.4100 33.2 34.3 n.a. 6.6 25.9100 20.6 38.1 n.a. 7.7 33.6

TABLE IV.1Rate fixation structure of new koruna credit to households(%)

Total < 1 year

1-5years

> 5 years

5-10years

> 10years

Source: CNB

Consumer credit

The global crisis on financial markets could lead to Czech banks tightening theircredit conditions above the yield curve, mainly because of uniform riskmanagement in the multinational banking groups to which Czech banks belong.An analysis of interest rates on new koruna loans showed that the interest rateconditions had tightened only slightly for the corporate sector. This tightening wassmaller than the increase in the corresponding yield curve rates. While three-monthinterbank PRIBOR rates rose by 103 basis points between June 2004 and January2008, average rates on new corporate loans with floating rates or fixations of lessthan one year increased only by around 50 basis points. However, some riskiersegments (small businesses and consumer credit to households) and also loans forhouse purchase saw some tightening above the yield curve. Yield curve rates withmaturities of over one year rose by about 40 basis points (rising more at the shortend), while rates on new loans for house purchase with fixations of between oneand five years increased by almost 100 basis points. At the same time, marketcontacts confirm that a number of segments recorded some tightening of non-interest lending conditions (required collateral, etc.).

New koruna deposits have also shown a gradual increase in rates in recent months.Average new client deposits were remunerated at 1.12% at the end of 2006,1.42% at the end of 2007 and 1.54% at the end of 2008 Q1. Growth is alsonoticeable for the generally higher rates on new time deposits. Such rates rose from2.05% at the end of 2006 to 2.57% a year later. They have continued increasingin 2008, reaching 2.63% at the end of March.

Although the Czech banking sector is dominated by foreign banks, such banks differin how they raise funds for credit expansion. While the three largest banks and thebuilding societies sector draw on wide deposit bases, medium-sized and smallerbanks can rely more on financing from foreign parent banks or the interbankmarket. An analysis of the tightening interest conditions on corporate loans by banktype reveals that large banks can profit from their relative financial independence offunding from foreign owners and thus increase interest rates less than medium-sizedand smaller banks (see Chart IV.13). A similar conclusion applies to loans for housepurchase, with the exception of fixations of up to one year (see Chart IV.14). Asregards the volume of new loans, an analysis of the data indicates that the share ofmedium-sized and small banks in total new loans declined slightly in all segments.

Profit and capitalThe banking sector generated a record net profit of CZK 47.1 billion in 2007, up by24% on 2006, which was also a successful year (see Chart IV.15). The main sourceof profit for most banks was growing income from financial activities. This was inline with the high return on equity (24.5%) and return on assets (1.3%) achievedin 2007.

The growth in profit from financial activities was driven primarily by interest profit,with annual growth of 19% and an almost 64% share of the total profit fromfinancial activities. The ratio of interest profit to non-interest profit has long beenaround 3:2 in the Czech Republic. The most important share of interest incomecomes from client loans. In some EU countries, especially the original members, theshare of non-interest profit has risen gradually in recent years. In some countries,this component has now exceeded interest profit, which is being depressed by lowrates and strong competition.67


67 In 2006, the share of non-interest profit exceeded 50% in Belgium, France, Germany and Luxembourg,i.e. countries with large banking sectors that significantly affect EU-wide aggregates due to their highweight in the total. These issues and international comparisons in other areas are addressed in moredetail in Davidová, P., Komárková, E.: Český bankovní sektor vs. evropské banky (The Czech BankingSector versus European Banks), Bankovnictví 2/2008.

CHART IV.13Change in interest rates on new corporate loans tohouseholds versus change in interbank rates(percentage points)

0.77 0.81

1.03

0.78

0.48

0.040.0

0.2

0.4

0.6

0.8

1.0

1.2

Interbank rates Small corporateloans

Large corporateloans

3M PRIBOR 1Y PRIBOR Large banks Medium-sized banks

Source: CNBNote: Change between June 2007 and January 2008; corporate loans with floating ratesor fixations of less than one year, covering around 90% of all new corporate loans.

CHART IV.14Change in interest rates on new house purchase loans tohouseholds versus change in interbank rates(percentage points)

1.161.01

0.51

1.06 1.07

1.30

0.35 0.40

0.13

1.03

0.78

0.190.04

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

Interbankrates

Fixation< 1 year

Fixation1-5 years

Fixation> 10 years

3M PRIBOR 1Y PRIBOR 3Y IRS 5Y IRSLarge banks Medium-sized banks Building societies

Source: CNBNote: Change between June 2007 and January 2008; the given fixations cover around95% of all loans for house purchase.

CHART IV.15Year-on-year growth in profit from financial and operating activities and net profit(%)

-40

-20

0

20

40

60

80

2001 2002 2003 2004 2005 2006 2007

Net profit/loss Profit from financialand operating activities

Source: CNB


Given the current slowdown in bank lending, interest profit growth can be expectedto moderate and its share to decline gradually in the Czech Republic as well. Banksnow operate in a competitive environment (especially as regards loans to households)that does not allow them to set their interest rates and fees above a reasonable level.

Profit generation is one of the most significant factors strengthening a bank's capitaland hence also the financial stability of individual institutions and the sector as awhole. The gradual decline in capital adequacy indicators, which can be observedsince 2003 (see Chart IV.16), is due to a combination of rising capital requirementsas a result of an expansion in lending and the start of a period of massive dividendpayments. At the end of 2007, capital adequacy and Tier 1 capital adequacy reached11.5% and 10.3% respectively. Both indicators suggest a sufficient level of capital.The slight increase in both indicators in 2007 was related mainly to high net profitgeneration and a year-on-year fall in dividends paid of 55%. Capital adequacygrowth was also aided by the gradual transition of Czech banks to the Basel IIframework, which, thanks to more accurate risk assessment, allows the banks thathave been governed by its rules since 1 July 2007 to set lower capital charges.

In the set of banks that switched to Basel II in 2007 H2 and were also required toreport reference values under Basel I (a total of five banks using the IRB approachto credit risk and having a share in the sector's assets of 48%), the capital chargedecreased by 13%−22% in individual months due to the new rules (see Table IV.2).All the banks recorded a decline. The Basel II rules are also more sensitive to thepossibility of adding/deducting individual items to/from regulatory capital. Four outof the five banks had to deduct missing provisions from capital as a result ofinsufficient creation of provisions, which fell short of the expected loss. Capital fellby 4%−5% in the group of monitored banks in individual months by comparisonwith the Basel I rules. As this decline was smaller than the fall in the capital charge,the resulting capital adequacy ratio under the applicable Basel II Decree was higher than the Basel I reference value. The bands of the capital adequacy ratios for the individual banks calculated under the two approaches partly overlap (seeChart IV.17).

The expected decline in the capital charge and capital cushion materialised in allfive banks. In the coming period it will be necessary to assess regularly whethermodels and processes are set correctly in all banks and whether the capital savingsachieved thanks to the new framework correspond to the risk profiles of theindividual banks.

The probability of default (PD) and loss given default (LGD) are important indicatorsunder Basel II. Based on data from the five Czech banks that introduced the Basel IIIRB approach in mid-2007, the average LGD was around 42%. This parameter wasthe same for exposures to both the corporate sector and the household sector. Ifthis value and the average default rate based on aggregated data from creditregisters68 were applied to exposures to households and corporations for the wholebanking sector in 2007, the aggregate capital charge for the whole banking sectorunder the Basel II IRB approach could be calculated. The baseline scenario based onthe CNB's official macroeconomic forecast (see sections 2.1 and 4.2) implies a slightrise in the default rate for both corporations and households.69 In line with the IBR

51

68 The historical 12-month default rate of non-financial corporations was calculated using aggregate datafrom the Central Credit Register, which is administered by the CNB and covers the whole banking sector. For the household sector, the default rate was estimated using data from the Banking ClientInformation Register operated by the Czech Banking Credit Bureau, which contains data for most banks.

69 Macroeconomic credit risk models for corporations and households predict an increase in the defaultrate of around 1.5 percentage points for corporations and around 0.5 percentage point for householdsat the end of 2008.

CHART IV.16Capital adequacy (%)

89

10111213141516

2000 2001 2002 2003 2004 2005 2006 2007

Capital adequacy Tier 1 capital adequacy

Source: CNB

CHART IV.17Capital adequacy ratios under applicable Basel II framework and reference Basel I framework(%, banks required to report under both frameworks)

5

10

15

20

25

30

7/07 8 9 10 11 12

Minimum under Basel I Maximum under Basel IMinimum under Basel II Maximum under Basel II

Source: CNB

CapitalCapital chargeCapital adequacy

CapitalCapital chargeCapital adequacy

105.0 104.7 104.8 104.8 95.2 100.880.1 78.9 80.6 82.7 84.1 87.110.5 10.6 10.4 10.1 9.1 9.3

Basel II

99.2 98.3 98.7 100.3 91.4 96.568.7 68.3 69.1 69.7 67.4 68.211.6 11.5 11.4 11.5 10.9 11.3

TABLE IV.2Selected indicators under Basel I and Basel II(CZK billions; %; banks required to report under both frameworks)

7/07 8/07 9/07 10/07 11/07 12/07Basel I

Source: CNB

approach under Basel II, the regulatory capital charge should therefore increase.While the capital charge for exposures to households should rise by only 2%, theexpected increase for exposures to corporations is 10%. The available evidencesuggests a negative relationship between the probability of default and economicgrowth.70 Thus, the capital requirement usually increases in an economic downturn,which may lead to restricted lending above all to the corporate sector, furtherexacerbating the decline in economic performance. However, this effect, oftenreferred to as the pro-cyclicality of Basel II in the literature, is dampened by asufficient capital cushion exceeding the mandatory capital charges.

The Basel II capital framework was gradually introduced in all EU countries in 2007(and at the start of 2008). The evaluation of the impact of the new rules is beinghampered by the effects of the US subprime mortgage crisis on European banks.Ultimately, the crisis led to significant losses of several large banks, resulting, amongother things, in a decrease in their capital adequacy.71

Building societiesAs in the case of bank loans to households, building society loans rose by roughlyone-third in 2007, amounting to CZK 180 billion, or 10% of total banking sectorlending. As in previous years, bridging loans were the biggest contributor to thisgrowth. Consequently, the share of bridging loans, which are to some extent analternative to bank mortgage loans, in total building society loans approached80%. Deposits with building societies grew by 7% year on year to CZK 394 billionin 2007, accounting for almost 20% of total deposits in the banking sector (seeChart IV.18). The ratio of loans to deposits rose to 45% at the end of the year.Building societies currently register around 4.8 million contracts with statecontributions and the market can be regarded as almost saturated. Over the entireexistence of the system, new loans amounting to CZK 28 billion have beenprovided on average each year. New standard building society loans have so farbeen fluctuating around several billion koruna a year.

The smooth development of the building savings system depends largely on theparameters of state support, which amounted to roughly CZK 15 billion in 2007.Over the entire lifetime of the building savings system, accumulated state supportof CZK 122 billion, or an average of CZK 10 billion every year, has been provided(see Chart IV.19). The current level of state support in the Czech Republic is high byinternational standards.72 Moreover, the building savings legislation does not allowa significant decrease in public expenditure to be achieved in the short run in theevent of a decision to change the parameters of state support. For example, achange in the size of state support would fully manifest itself at a horizon of morethan six years. Under the current legislation, building societies can allow theirclients to increase the target amount, thereby prolonging the original contracts andletting them draw on state support even after the expiration of the compulsorysaving period. Table IV.3 shows that more than 800,000 prolonged contracts withthe maximum state contribution of CZK 4,500 existed within the system at the endof the year.

The stability of building societies under the existing system depends on theexistence of a significant proportion of "friendly clients" (drawing state support


70 Jakubík, P. (2007): The Macroeconomic Environment and Credit Risk. Czech Journal of Economics andFinance, 1-2/2007, pp. 60−78.

71 Financial Stability Review, ECB, June 2008. 72 In 2007, state support amounted to CZK 3,000 (EUR 112) in the Czech Republic, EUR 88 in Germany

(consisting of a housing bonus of EUR 45 and a contribution to the employee saving scheme of EUR 43), EUR 35 at most in Austria and EUR 60 in Slovakia.

CHART IV.18Building society loans and deposits(CZK billions; %)

0

100

200

300

400

19971998199920002001200220032004200520062007020406080100

Building society loans, total (CZK billions)Bridging and mortgage loans (CZK billions)Building society deposits (CZK billions)Loans to clients/Building society deposits (right-hand scale; %)Bridging loans/Total loans (right-hand scale; %)Standard loans/Total loans (right-hand scale; %)

Source: CNB, MF CR

CHART IV.19State support for building savings schemes(CZK billions; ‰)

010203040506070

1997 1998 19992000 2001 20022003 20042005 2006 20070

5

10

15

20

25

Standard building society loans (stock)State support (annual expenditure from state budget)New loans (standard and bridging)Annual state contribution to contract/Target amount(right-hand scale; ‰)

Source: CNB, MF CRNote: State support estimated for 2007.

a) Old contracts without prolongation (until 31 Dec. 2003)of which old contracts withoutstate support entitlement

b) Prolonged contracts with entitlement of up to CZK 4,500

c) New contracts (since 1 Jan. 2004) with entitlement of up to CZK 3,000of which new contracts without statesupport entitlement

Contract total − with state support entitlement

− without state supportentitlement

Situation as of 31 December 2007

2,629 239 2.55 207

256 26 x x

858 97 2.19 352

1,609 40 1.92 263

64 2 x x

4,776 348 2.39 249

922 28 x x

TABLE IV.3Overview of building savings system

No. ofcontracts

thous.

SavedamountCZK bn

Av. interestrate on

deposits% p.a.

Av. target

amountCZK

thous.

Source: CNB


without investing in housing) and may be adversely affected by parametric changesin state support for building savings, interest rate movements and changes in otherfinancial market segments and the state's policies towards them. Building societiesface a relatively high level of interest rate risk, as they offer fixed contractual ratesand have a preponderance of saving clients, or, in other words, an excess ofdeposits over loans. Over the course of 2007, building societies managed to reducetheir interest rate risk by cutting their deposit rates well below 3% in the contractprolongation process. However, only an increase in the loan-to-deposit ratio canbring about a further decline in interest rate risk.

4.1.2 Non-banking financial institutions

Insurance companiesThe most important categories on the insurance market are the traditionalsegments of life and non-life insurance. Life insurance includes permanent lifeinsurance and combined term and permanent insurance (46% of life insurancepremiums). Life insurance combined with an investment fund (unit-linked) isgaining popularity (34%). As regards non-life insurance, the most importantcategories are vehicle liability insurance (30% of non-life insurance premiums),property insurance for entrepreneurs and private individuals (22%), vehicle accidentinsurance for entrepreneurs and private individuals (21%) and business insurance(20%).

Premiums written rose by 8.9% year on year in 2007. This growth was driven bylife insurance (14.6%), especially investment (unit-linked) life insurance. Premiumswritten in non-life insurance were up by 5.2% (see Chart IV.20).

Previous analyses and international comparisons73 reveal that there is room on theCzech insurance market for further growth over the long run, especially as regardspremiums written and financial placement (as a percentage of GDP) in both life andnon-life insurance. The ratio of financial investment in non-life insurance to GDPincreased from 2.1% to 2.6% year on year. Given the rising volatility of climatechange, a gradual reassessment of non-life insurance contracts and a rise inpremiums under updated insurance schemes can be expected.

Insurance companies are creating higher technical provisions for life insurance.Claim settlement costs in life insurance rose from 33% to 41% of total claimsettlement costs in 2007. Within non-life insurance, vehicle liability (17%), vehicleaccidents (16%) and natural disasters (14%) accounted for most of the claimsettlement costs. Non-life insurance segments have faced repeated shocks andclaim settlement costs, mainly in connection with damage caused by the floods in2002 and the hurricane in 2007 (see Chart IV.21). In non-life insurance, the ratio ofclaim settlement costs to technical provisions was usually higher than in lifeinsurance. Premiums written were also higher due to a shorter claim settlementcycle (see Chart IV.22). Selected non-life segments (vehicle accidents, naturaldisasters, damage to property) require the involvement of reinsurers. Reinsurersaccounted for 15% of total claim settlement costs, of which 1% in life insuranceand 25% in non-life insurance.

Technical provisions are a source of funds for investment in financial assets.Insurance companies invested 50% of their funds in bonds of banks andinternational institutions and 6% in reinsurance companies. Other investmentswere made in mortgage bonds and mutual fund units, property and marketableshares and bonds (see Chart IV.23).

53

73 Financial Stability Report 2006, CNB.

CHART IV.20Life and non-life insurance (premiums written) (CZK billions)

020406080

100120140

1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

Insurance, total Life insurance Non-life insurance

Source: CNB

CHART IV.21Claim settlement costs(CZK billions)

0

20

40

60

80

1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

Non-life insurance Life insurance

Source: CNB

CHART IV.22Cost ratios of main insurance segments(coeff.)

C / TP (gross)

0.00.20.40.60.81.01.21.4

0.0 0.2 0.4 0.6 0.8 1.0 1.2

C / Premiums

Non-life Life

Source: CNBNote: Claim settlement costs (C), technical provisions (TP) and premiums written are atgross value (not adjusted for the effect of reinsurers).

CHART IV.23Financial investment in assets(% of financial investments)

50.5%

17.5%

15.5%

2.7%5.5%

2.0%6.3%

Bonds of central banks, other banks and international institutionsOther bonds, bills, deposit certificatesMortgage bonds and mutual fund unitsForeign securities tradedShares and propertyCredit, loans and other claimsClaims on reinsurers

Source: CNB

Insurance companies were compliant with the solvency criteria (according toaudited 2006 results), as their internal funds were greater than or equal to therequired solvency margin (100%). Under the existing legislation, the aggregateavailable margin was three times the required solvency margin on the life insurancemarket and 3.3 times that on the non-life insurance market.

Under the planned Solvency II framework,74 a total of 12 domestic insurancecompanies underwent a solvency calculation under the quantitative impact study(QIS3). The insurance companies had to meet the capital requirements under thenew rules − the solvency capital requirement (SCR) in the case of mobilisation ofown funds for some of them and the minimum capital requirement (MCR). A totalof ten insurance companies met the solvency criterion under the methodology usedwithout additional capital needs. As in the case of stress testing (see section 4.2), itwas found that capital requirements for market risks (especially interest rate andequity risks) were most important for life insurance, while the capital requirementfor non-life insurance risk was predominant in non-life insurance.75

Insurance company stability was fostered by return on equity, which reached21.7% in 2007, and average return on assets, which was 3.7%. Strongly risingcontract acquisition costs acted in the opposite direction. These costs could place aburden on the insurance system in the future (see Table IV.4).

Pension fundsAt the end of 2007, a total of CZK 162.4 billion in contributions was registered onthe accounts of private pension planholders. State contributions accounted for CZK 21.9 billion of this amount. Funds from employers totalling CZK 16.6 billion,to which the state contribution does not apply, receive preferential tax treatment.Overall, CZK 73.3 billion has been paid in benefits since 1994, of which CZK 48.8 billion as lump-sum settlement and CZK 8 billion as terminationsettlement. The other items paid include retirement, service, survivors' and disabilitypensions and other payments.

Contributions from planholders have recently been rising (by 14.2% year on yearin 2007). This growth in funds has been supported by the state contribution andtax deductions and by a stronger motivation among individuals to provide for theirold age. Benefits paid have also risen quite strongly in the last two years (see Chart IV.24).

The ratio of pension insurance to GDP was 4.7% in the Czech Republic in 2007, a low figure by comparison with the data available for selected countries (see Chart IV.25). Nevertheless, the number of planholders has increased significantly to 3,936,000 since 2005 (the rates of increase were 11.3% in 2005, 10% in 2006and 9.5% in 2007). Planholders represent more than one-third of the Czechpopulation.


74 The new Solvency II regulatory framework in the insurance industry affects not only the technical riskfor life, non-life and health insurance, but also market risks and credit and operational risk. Under aprocedure organised by CEIOPS, the QIS3 results for the calibration of the standard formula for calculation of the minimum capital requirement (MCR) and the solvency capital requirement (SCR) were published in November 2007. In 2008, a QIS4 study is under way to recalibrate the capitalrequirements, technical provisions and other processes in solvency calculation.

75 CEIOPS' Report on its Third Quantitative Impact Study (QIS3) for Solvency II, CEIOPS, October 2007. Theresults of the study for the Czech insurance market are described in Justová, I., Kotaška, M. (2007):Vyhodnocení výsledků třetího kola kvantitativní dopadové studie (QIS3) za český pojistný trh(Assessment of the Results of the Third Round of the Qualitative Impact Study (QIS3) for the CzechInsurance Market). Pojistný obzor 12/2007.

1. Contract costs for payment in given yearYear-on-year growthShare in profit after taxation

2. Contract acquisition costs as prepaymentsYear-on-year growthShare in profit after taxation

Ratio of costs (1.+2.) to annual premiums writtenRatio of costs (1.+2.) to total costsCoverage of total costs by annual premiums (in years)

17.9 8.4159.8 121.2

43.6 22.943.1 26.919.4 17.28.5 8.0

2.3 2.2

TABLE IV.4Insurance contract costs paid to intermediaries(%)

2007 2006

Source: CNB

CHART IV.24Pension fund sources and amounts disbursed in given year (CZK billions)

05

101520253035

1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

State contributionsPlanholders' contributions (including advances since 2005)Contributions paid by employer(including planholders' contributions)Amounts disbursed

Source: CNB

CHART IV.25Pension planholders' funds(% of GDP)

NL UK

IE DKPT PL ES

AT

LVSKIT

EESICZBE

0

50

100

150

0

2

4

6

Countries with higher ratio to GDP (left-hand scale)Countries with lower ratio to GDP (right-hand scale)

HU

Source: CNB, ECBNote: Data for 2006. Selected EU countries.


As in the insurance industry, efforts to win clients led to an increase in contractacquisition costs (including prepayments) in 2005−2007. Intermediaries also madeuse of the possibility of transferring money to other funds and charged higheramounts for new contracts than in the past. Planholders' contributions rose by14% year on year, but contract intermediation costs increased by 30% in the 2007result and another 20% in prepayments. The increase in costs may place a burdenon the pension scheme industry in the future (see Table IV.5).

Under the limits set by law, pension funds should invest the funds they raise fromplanholders in relatively safe assets. At the end of 2007, 84.8% of assets wereinvested in bonds issued by general government, deposits with domestic banks andother bonds (see Chart IV.26). 10.4% of funds were invested in shares and units,which can be more volatile. Investments in shares and mutual fund units were lessfavourable in terms of profit generation. In 2007, funds recorded losses of 6.6% ofthe acquisition price from investments in shares and 12.5% from investments inmutual funds, i.e. a total of CZK 1.5 billion. The value of funds' assets was adverselyaffected by the appreciating koruna and rising market interest rates.76 According totheir balance sheets, annual valuation losses in the total assets of pension fundswere CZK 5.6 billion. Profit of pension funds was CZK 4.4 billion. The realperformance of their assets was slightly negative as a result of the valuation losses.

The existing pension funds are designed to provide a non-negative annual yieldwhich, after coverage of fund administration costs, should ensure that the client'scontribution gains in value owing to continuous efforts by the manager to raise thevalue of the fund's assets in real terms.77 However, growing volatility on the assetmarket and in particular negative differences between the real value of assets andtheir acquisition price may have an adverse effect on the funds' assets in the longerrun, especially if benefits rise or are paid at a faster pace (primarily lump-sumsettlements). Commensurate capital increases by funds' shareholders would bedesirable as protection against this risk (see Table IV.5).

Investment companies and mutual fundsIn 2007, a total of 18 investment companies were operating on the capital market,three of which were controlled by resident banks. The companies maintained ahigh RoE of 46.5% and RoA of 25.6%. Investment companies usually administerdomestic open-ended mutual funds. The accounts and transactions of these fundsare separate from transactions for the company's own account.

Domestic open-ended mutual funds are a form of collective investment designedmainly for individual investors. At the end of 2007, 120 funds, with assets of CZK 191 billion, were active. Of this number, 10 funds, with assets of CZK 77 billion, were money market funds (see Chart IV.27). A total of 66 funds,with assets of CZK 146 billion, were administered through domestic subsidiarybanks. Thus, banks contribute significantly to the intermediation of transactionsand offer units as an alternative to bank deposits.

55

76 For example, unsecured currency exposures of funds would result in valuation losses of around CZK 1.5 billion if the koruna appreciated by 10%. Further losses would be caused by a fall in market pricesof bonds. If interest rates rose by 1%, the valuation losses would be CZK 2.5 billion (see also section 4.2).

77 In recent years the possibility of separating the accounting of planholders' assets from that of the assetsof the pension fund's shareholder (manager) has been discussed. The World Bank published a study onpension funds in the Czech Republic: Pilot Diagnostic Review of Governance of the SupplementaryPrivate Pension Fund Sector, The World Bank, January 2007. This evaluation study, prepared at therequest of the Ministry of Finance, focused on principles of management in the supplementary pensions sector and should provide recommendations leading to better management and control in thesector and enhanced protection of planholders.

1. Contract costs for payment in given yearYear-on-year growthShare in profit from financial operationsShare in profit after taxation

2. Contract acquisition costs as prepaymentsYear-on-year growthShare in profit after taxation

Ratio of costs (1.+2.) to annual state supportValuation differences between acquisition price and fair value of assets 1/

Share in pension funds' capitalRatio to annual state support

30.2 20.261.6 36.021.2 17.4

20.2 29.578.5 69.695.8 88.9

-124.3 38.3-96.2 28.7

TABLE IV.5Pension scheme contract costs paid to intermediaries andeffect of asset revaluation(%)

2007 2006

Source: CNBNote: 1/ A negative value means that the fair (market) value fell below the acquisition priceof the assets.

CHART IV.26Structure of pension fund investments(%; 2007)

58.6%

3.4%

13.2%

11.1%

9.6%4.1%

Bonds issued in Czech RepublicBonds of OECD member states and international institutionsOther bonds tradedShares, units and real estateDeposits on bank accountsOther assets

Source: CNB

CHART IV.27Assets of domestic open-ended mutual funds(CZK billions)

0

50

100

150

200

1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

Equity BondMoney market MixedFunds of funds OtherDomestic OMFs, total

Source: CNB, AFAM CR

Foreign mutual funds offer products on the domestic market through registeredinvestment intermediaries and investment firms (see Chart IV.28). The investmentitself is carried out by investment companies (funds) registered abroad. The totalinvested in foreign funds in the Czech Republic in 2007 was CZK 142 billion, 90%of which was intermediated by banks (see Chart IV.29).

As regards the structure of financial investment, bond funds and money marketfunds lost market share. Clients showed interest in mixed funds and funds of funds.The first investments in new domestic real estate funds were also recorded. Therewas growing interest in guaranteed funds offered from abroad, which offer acontractual guarantee of return of principal and a minimum yield. The investmentsin these funds were the highest, almost twice as high as investments in foreignequity funds.

Investment firmsThere were 44 investment firms active in the capital market at the end of 2007, ofwhich 31 were non-banks.78 Total assets of non-bank investment firms reachedaround CZK 25 billion in 2007, a moderate increase of 20% compared to theprevious year. Loans and other receivables accounted for the largest part (80%) of their assets, owing to the nature of their business. Their profit totalled CZK 890 million at the end of the year. Non-bank investment firms generallyachieved higher profitability (RoE 22.2% and RoA 3.6%). The capital ratio of non-bank investment firms was high (average more than 200%, median 25%), but thevalues were dispersed over a wide range (from 8% through to 5,000%). Non-bankinvestment firms administered client assets of CZK 393 billion in their balancesheets at the end of the year. As regards the structure of active clients, whosenumber exceeded 31,000 at the end of the year, other clients prevailed in bothmanagement and other relationships, accounting for 40% (see Chart IV.30). Giventhe nature of the business of non-bank investment firms (frequent changes intrading portfolio positions) and their predominant focus on the Czech financialmarket, the risk of the current crisis affecting this segment is rather low.

Non-bank financial corporations engaged in lendingThere were 223 leasing companies (with assets of CZK 278 billion), 57 otherlending companies (with assets totalling CZK 100 billion) and 10 factoring andforfaiting companies (with assets of CZK 24 billion) active on the non-bank credit market at the end of 2007. CZK 222 billion was lent in leasing, of which CZK 143 billion to corporations and CZK 77 billion to households (see Table IV.6).Annual growth in leasing loans (5.3%) had been low by comparison with bankloans (19.9%) in 2006, but these indicators were almost identical at 17% last year.This was probably due to expected changes in the tax reform, which is reducing thetax breaks offered on financial leasing. These tax changes seem to have led to astocking-up effect. Although leasing loans are used mainly by non-financialcorporations (65%), annual growth in such loans was higher for households (34%).Consumer credit, hire-purchase loans and credit card loans from other lendingcompanies totalled CZK 75 billion, the overwhelming majority of which wasprovided to households. However, their annual growth of 20% was below that ofbank consumer credit provided to households (35.1%). Total loans to non-financial


78 Nine non-bank investment firms are members of the Prague Stock Exchange and the volume of theirtrading was CZK 1,456 billion in shares and CZK 2.97 billion in bonds in 2007. While their share trading increased by 20%, their trading in bonds fell by almost one-half compared to 2006. Bank and non-bank investment firms carried out transactions in the two types of instruments totallingCZK 1,998 billion.

CHART IV.30Structure of investment firm clients by volume of transactions arranged in 2007(%)

35.6%

12.7%0.3%11.9%

41.6%

Banks Institutional investorsCredit unions Investment firms

(non-banks)Others

Source: CNB

Leasing companiesLoans, totalLoans to non-financial corporationsLoans to householdsOther lending companiesLoans, totalLoans to non-financial corporationsLoans to householdsFactoring and forfaiting companiesLoans to non-financial corporations

182.6 190.2 222.1 16.8126.8 131.2 143.4 9.453.8 57.3 76.9 34.4

54.1 62.3 74.7 20.02.1 3.3 4.6 38.8

51.3 57.5 69.1 20.2

14.2 16.9 20.4 20.6

TABLE IV.6Activities of leasing companies, other lending companiesand factoring and forfaiting companies (CZK billions; year-on-year change in %)

2005 2006 2007 y-o-ychange

06/07

Source: CNB

CHART IV.29Assets of foreign open-ended mutual funds(CZK billions)

0

50

100

150

Equity Bond Moneymarket

Mixed Guaranteed ForeignOMFs, total

2005 2006 2007

Source: AFAM CR and AKAT

CHART IV.28Assets of open-ended mutual funds(CZK billions; %)

050

100150200250300350

12/02 6/03 12 6/04 12 6/05 12 6/06 12 6/07 12051015202530

Foreign OMFs (left-hand scale)Domestic OMFs (left-hand scale)Ratio of OMF assets to household deposits in banks(right-hand scale; %)

Source: CNB, AFAM CR


corporations backed by receivables were CZK 20 billion in 2007. Their rate ofgrowth picked up slightly, by 1.6 percentage point year on year.

A potential risk arising from the activities of non-bank credit institutions is the factthat they are not subject to direct supervision. Many leasing companies, however,are controlled by banks or other large financial institutions.

4.2 ASSESSMENT OF THE FINANCIAL SECTOR'S RESILIENCE

According to stress tests, the financial sector is currently resilient to market, creditand some specific risks. Only an extreme macroeconomic scenario with significantadverse impacts on interest rates, the exchange rate and GDP growth wouldnecessitate capital injections to ensure compliance with the regulatory limits andmaintain sufficient capitalisation in financial institutions. This is particularly true ofpension funds, which are very sensitive to market risks according to the tests. The banking sector stability indicator confirms a continuing process of capitaloptimisation in the banking sector, with unchanged resilience to the main risks.Tests of banks' balance-sheet liquidity indicate that the banking sector is sufficientlyresilient to an outflow of deposits and some other hypothetical changes in thefinancial market. However, only institutions with a strong deposit base withstandthe extreme variant of pressures on balance-sheet liquidity.

This section sets out to assess the resilience of the Czech financial sector. This isdone using stress tests quantifying the impacts of various shocks on financialinstitutions as well as some supplementary indicators. In the stress testing, weanalyse in particular the effects of alternative model-consistent scenarios. Sensitivityanalyses were also performed to assess some specific risks in more detail. Thissection also presents tests of the banking sector's balance-sheet liquidity for thefirst time (see Box 7).

The three alternative scenarios were introduced in sections 2 and 3 of this Report:"safe haven", "property market crisis" and "loss of confidence". All the scenarioswere defined primarily by the evolution of key macroeconomic indicators such asGDP, inflation, the unemployment rate, short-term interest rates and the exchangerate. They were prepared using the CNB's official forecasting model (see Table IV.7).The other parameters entering the stress tests were derived using the values ofthese macroeconomic variables with the aid of sub-models and expert estimatesbased on historical averages or foreign experience. The key parameter for credit risktesting, i.e. the ratio of non-performing loans to total loans, was generated usinga credit risk model and a credit growth model.79 Parameters from the asset markets,i.e. stock prices, long-term yields and real estate prices, were set by expertjudgement (see Tables IV.8 and IV.9).

The impacts of the individual alternative scenarios can be compared with the mostprobable path of the economy, as expressed in the baseline scenario. This is basedon the CNB's official February 2008 macroeconomic forecast and assumes aslowdown in GDP growth in 2008, gradual appreciation of the exchange rate anda higher inflation rate, which, however, will start to decline in Q2.80 As the baselineis not a shock scenario, the stress tests in this case use the average predicted valuesof the macroeconomic variables for 2008 (see Tables IV.8 and IV.9).

57

79 Both models are described in detail in the thematic article Credit Risk and Stress Testing of the BankingSector in the Czech Republic in the 2006 Financial Stability Report.

80 The baseline scenario is described in detail in the CNB's Inflation Report, I /2008, February 2008.

Real GDP growth (%; y-o-y)Inflation rate − CPI (%; y-o-y)Unemployment rate (%)1Y PRIBOR (%)CZK/EUR exchange rate

4.1 2.4 0.3 2.86.2 7.0 5.3 8.06.0 6.3 6.7 6.33.8 2.8 1.5 8.7

... 1/ 25.6 27.0 30.5

TABLE IV.7Calibration of baseline and alternative scenarios(2008 averages)

Baseline Scenario A Scenario B Scenario C

Source: CNBNote: 1/ In 2008, the baseline expects a correction of the record values initially and thena slight appreciation

Change in CZK interest ratesChange in EUR interest ratesChange in CZK/EUR exchangerate (− appreciation)Loan default rateTotal credit growthInterbank contagion risk"Change in property prices (+ rise, - fall)"

-0.2 p.p. 0.1 p.p. -0.9 p.p. 4.4 p.p. -0.8 p.p. -1.4 p.p. -0.4 p.p. -0.4 p.p.

- -6.7% -0.4% 20.1%4.2% 5.2% 6.9% 4.9%

16.4% 9.9% 14.6% 4.9%x x x x

15% 0% -30% -5%

TABLE IV.8Scenario type and shock size in bank stress test

Baseline Scenario A Scenario B Scenario CScenario type

Note: Changes in parameters represent the difference between 2007 Q4 and 2008 Q1,or, in the case of the baseline, between 2007 Q4 and the average for 2008.

Change in CZK interest ratesChange in EUR interest ratesChange in CZK/EUR exchangerate (− appreciation)Increase in default loans (reclassification)Change in share value Change in property prices (+ rise, - fall)"Increase in LI1)

(risk of epidemics)Increase in NLI1)

(risk of climate change)

-0.2 p.p. 0.1 p.p. -0.9 p.p. 4.4 p.p. -0.8 p.p. -1.4 p.p. -0.4 p.p. -0.4 p.p.

- -6.7% -0.4% 20.1%

4.2% 5.2% 6.9% 4.9%0% -15% -15% -15%

15% 0% -30% -5%

3% 3% 3% 3%

50% 50% 50% 50%

TABLE IV.9Scenario type and shock size in insurance company andpension fund stress test


Note: 1/ Insurance company test only.Note: Changes in parameters represent the difference between 2007 Q4 and 2008 Q1,or, in the case of the baseline, between 2007 Q4 and the average for 2008.

The overall effects of the shocks − also referred to as losses − are expressed in bothabsolute and relative terms (relative to capital or past profits).81 It is also assumedthat the individual financial institutions generate profits at the level of the averageof the last few years, which could be used to cover the losses resulting from theshocks.82 The allocation of profit to cover losses would continue until the originalcapital adequacy ratio was reached again, provided that the volume of profit allowsthis.

The methodology and results of the stress tests for banks were described in theprevious Financial Stability Reports. Compared to last year, the calculation ofinterest rate risk by currency was refined (detailed calculation for the bond portfolioin koruna and in foreign currency) and the effect of a decline in real estate priceswas included for the first time.

The CNB first published the methodology and results of stress tests for insurancecompanies and pension funds in its 2006 Financial Stability Report.83 The insurancecompany stress tests apply the currently valid solvency calculation and capitaladequacy conversion to allow comparison of the results with banks. The capitalrequirements relate to equity capital and are calculated separately for life and non-life insurance. Stress testing of insurance companies captures the same marketrisks84 and credit risk (or risk of counterparty default) resulting from the individualscenarios as in the case of the banking sector. Moreover, the test for insurancecompanies includes shocks specific to the insurance sector. In the case of non-lifeinsurance the shock concerns climate change and the property consequences ofnatural disasters (the risk of catastrophic events). The hypothetical shock to lifeinsurance is associated with the risk of occurrence and consequences of epidemics.The life insurance shock was defined as an increase in premium reserves, premiumswritten or gross technical provisions depending on the category of life insurance.These items were in all cases increased by 3%. In non-life insurance, the shock forthe scenario was set as a 50% rise in gross claim settlement costs in a reference(usually three-year) period and was derived from historical experience (the insuredlosses during the floods in the Czech Republic in 2002). Where current reserves orpayments by reinsurance companies are not sufficient for insurance companies andthe required solvency margin would fall, the uncovered part of the effect of theshock represents the capital requirement. The tests take into account theparticipation of reinsurers within the scope laid down for the solvency calculation.85


81 These are "gross" losses representing the overall impact of the shocks. The banking sector as a wholecan continue to generate profit even in the event of such gross losses if it is able to generate sufficientincome to cover them. The stress tests incorporate the response of financial institutions to the negativeeffects of the shocks. It is therefore assumed that financial institutions will use profits (or the incomegenerating them) as a first line of defence against a drop in capital adequacy.

82 This is a relatively optimistic assumption, as weaker demand for financial services can be expected inaddition to a decline in the value of asset holdings and higher costs due to the fall in quality of the loanportfolio in the event of adverse economic developments. This would lead to a slowdown in activity offinancial institutions, affecting both interest and non-interest income.

83 Central banks and authorities supervising insurance companies in numerous EU countries are currentlyengaged in stress testing of insurance companies. Similar tests are also part of the InternationalMonetary Fund and World Bank's reports under the Financial Sector Assessment Program (FSAP). The International Association of Insurance Supervisors has issued recommendations for stress testing ofindividual insurance companies.

84 The calculation of the effect of the exchange rate shock in the test should be viewed as approximate,since it is based on only partial information on foreign currency assets and instruments. However, relatively low foreign currency liabilities exist in the balance sheets of insurance companies.

85 The effect of shocks on insurance companies' claim settlement costs could be fully transferred to reinsurers. In the Solvency I stress testing, the calculation of the required solvency margin includes minimum coverage by insurance companies themselves of 50% for non-life insurance and 85% for lifeinsurance.

CHART IV.31Results of macro stress testing scenarios

02468

101214161820222426

2006 2007 2008

Capital adequacy(%)

Share of default loans(%)

Growth of total loans


(%)

Source: CNBNote: Growth in total loans is defined as the average annual rate of growth. The share ofdefault loans relates to the estimation of the loan volume at the end of 2007.


The pension fund tests include market risks (interest rate, equity and exchange raterisk) and counterparty default risk (credit risk) and are similar to the tests applied toinsurance companies.

Interest rate risk is the most important market risk in the pension funds segment.Unlike last year, when it was based on aggregate data, the calculation of the effectof an interest rate shock is based on detailed data on the debt instrument portfoliosof the individual institutions. These data (including the currency, maturity andcoupon of the individual instruments) enable very precise calculation of theportfolio's value in the event of interest rate changes. Separate koruna and foreigncurrency portfolios were used to analyse the impacts of the alternative scenarios.

Impact of alternative scenarios on the banking sectorThe impact of scenario A ("safe haven") on the banking sector is relativelymoderate compared to the effects of the other scenarios (see Table IV.10). The totaleffects of the shocks would be CZK 53 billion (roughly 24% of the banks' capital),or 112% of the average annual profit in the last five years. Under this scenario,losses in the banking sector would be driven by an increase in credit risk, especiallyfor non-financial corporations (around 53% of the total losses). This is due to thestrong koruna, which would cause problems with loan repayments in the export-oriented corporate sector. However, the decline in GDP growth would also bereflected in households' income and their ability to repay their obligations. Thedefault rate for the overall portfolio would rise from 2.8% to 5.8%. This would bea combination of an increase from 3% to 7.9% for corporations and a rise from2.7% to 3.1% for households. As regards market risks, certain losses would besuffered in the event of currency appreciation. Any interbank contagion wouldresult in further − albeit limited − losses. After the allocation of disposable profit,the capital adequacy ratio would fall from 11.5% to 11%, remaining high abovethe regulatory minimum (see Chart IV.31).

The impact of scenario B ("property market crisis") is slightly higher than that of theprevious scenario. The overall effects amount to CZK 57 billion (120% of averageprofit). Capital adequacy would decline from 11.5% to 10.8%, also remaining highabove the regulatory minimum (see Chart IV.31). Under scenario B, the losseswould again be driven mainly by credit risk vis-a-vis non-financial corporations.However, greater defaults would also be recorded in the household sector owingto the decline in property prices. The aggregate default rate would rise to roughly7.1% in 2008.

Scenario C ("loss of confidence") would have the strongest impact on bankingsector stability. This is an extreme scenario, with the total effects of the shocksreaching CZK 118 billion, or 250% of the average profit in recent years. A capitalinjection of CZK 38 billion would be needed to keep capital adequacy at theregulatory minimum. The losses would be driven by interest rate and credit risks.Owing to long currency exposures, the strong depreciation of the koruna wouldnot cause losses, but the upward pressure of the weak currency on prices wouldlead to high interest rates and losses due mainly to a decrease in bond prices. Thehigh interest rates and a decline in GDP would result in a rise in the default rate forcorporations and households to around 5.8% in 2008.

Impact of alternative scenarios on the insurance sectorAll three scenarios would cause negative effects in the insurance sector, rangingfrom CZK 9.8 billion (around 85% of the average profit in the last two years) underscenario A to CZK 6.1 billion (54% of profit) under scenario B and a relativelystrong impact of CZK 19 billion (167% of profit) under scenario C, due to marketshocks, credit risk and the specific risks tested (see Table IV.11). To balance anylosses, insurance companies would use pre-tax profits (which are assumed to reachthe average of the last two years in the absence of shocks) and equalisation

59

Capital adequacy (CAR) 1/

Results for chosen scenario typeOverall impact of shocks (p.p. CAR)

Interest rate shockExchange rate shockCredit shock… households… non-financial corporations Interbank contagion 2/

CAR before profit allocationProfit allocation (p.p. CAR) 3/

Post-shock CARCapital injection (% of GDP) 4/

Share of banks with negative capital after shock 5/

11.5 11.5 11.5 11.5

-2.1 -2.8 -3.0 -6.30.2 0.1 0.6 -2.6

-0.1 -0.2 0.0 0.5-2.0 -2.4 -3.3 -3.6-0.5 -0.5 -0.5 -0.5-1.0 -1.5 -2.0 -0.6-0.2 -0.2 -0.2 -0.79.4 8.7 8.5 5.21.8 2.3 2.2 2.8

11.3 11.0 10.8 8.10.0 0.1 0.1 1.1

0.0 0.0 0.0 14.9

TABLE IV.10Results of bank stress tests(capital adequacy; % and p.p.)

Baseline Scenario A Scenario B Scenario C2007 2007 2007 2007

Scenario type

Notes:1) CAR means the capital adequacy ratio defined in accordance with the relevant CNB

regulations (in particular those governing the capital adequacy of banks and other prudential business rules).

2) Test integrated with interbank contagion and expected level of loss given default (LGD)100% and chosen probability of the banks' failure (default) on the basis of the CAR.

3) The scenarios assume that in the absence of shocks each bank would generate profit(or loss) equal to the average for the previous five years and that it would use any profit (income) as a first line of defence against a declining CAR.

4) The capital needed to ensure that each bank has a post-shock CAR of at least 8%.5) Market share of banks with negative capital after the impact of the assumed shocks

(as a percentage of total assets).

CAR 1/ for insurers as a whole (%)Overall impact of shocks from exposures (p.p.)

Interest rate shockExchange rate shockCredit shockEquity shockProperty price shock

Overall impact of shocks in insurance (p.p.)

Life insuranceNon-life insurance…motor vehicle insurance...climate change, natural disasters, property

CAR before allocation of profit and eq. provisions Allocation of profit and equalisation provisions (p.p.)

Post-shock CAR (%)Capital injection (% of GDP)

13,3 13,3 13,3 13,3

1.2 -3.0 -1.7 -6.31.3 -0.1 1.3 -5.7

-0.3 -0.6 0.0 1.9-0.2 -0.2 -0.3 -0.40.0 -2.1 -2.1 -2.10.3 0.0 -0.7 -0.1

-0.4 -0.5 -0.4 -0.5-0.1 -0.1 -0.1 -0.1-0.4 -0.4 -0.4 -0.4-0.2 -0.2 -0.2 -0.2

-0.2 -0.2 -0.2 -0.2

14.1 9.9 11.2 6.5

-1.3 2.8 1.6 3.912.8 12.7 12.8 10.40.2 0.2 0.2 0.3

TABLE IV.11Results of insurance company stress tests(capital adequacy; % and p.p.)

Baseline Scenario A Scenario B Scenario C2007 2007 2007 2007

Scenario type

Note: 1) Calculation for June 2007, derived for illustration from bank capital adequacymethodology in 2006.

provisions (in non-life insurance), if available, to prevent a decline in solvency in lifeand non-life insurance below the set minimum of 100% or to prevent a decline incapital adequacy.

Insurance companies would be able to withstand the extreme stress ensuing fromspecific shocks (climate change, epidemics) with an impact of CZK 60 billion insettlement costs at a horizon of three or more years (even though this figureexceeds the clean-up costs of the 2002 floods by roughly one-third). This is thanksto the volume of technical provisions, the spreading of claim payments over severalyears, the involvement of, and payments by, reinsurance companies, and partpayment of losses by policyholders. Insurance companies are capable of respondingto an increasing frequency of climate change manifestations and potential risinglosses and claim costs by changing their procedures, particularly in non-lifeinsurance. Not even the extreme scenario C combined with the other shocks wouldcause solvency to drop below the minimum of 100%.

The stress test results indicate that the insurance company sector as a whole wouldbe able to withstand even relatively strong shocks, taking into account currentcapitalisation, technical provisions and the utilisation of collateral. The worst resultwas recorded under scenario C, with the strong interest rate and equity shockmanifesting itself most in life insurance provided by universal insurers. The extentof the impact of the combination of adverse shocks is clearly visible in the solvencyratio before allocation of profit and equalisation provisions (see Table IV.12).

Impact of alternative scenarios on the pension fund sectorThe stress testing of pension funds suggests that they are highly sensitive tounfavourable economic developments and do not always have a sufficient capitalcushion to cover market risks. Given the high sensitivity to interest rate risk, thelargest losses would be recorded if scenario C, which assumes a surge in interestrates, were to materialise (see Table IV.13). The aggregate effect of the shocks onthe funds would be very strong if this shift in the yield curve were to beaccompanied by currency appreciation and a decline in the value of share holdings.The losses under scenario C would run to 175% of last year's (average) profit andwould require relatively extensive capital injections.

In addition to interest rate and equity risks, pension funds are relatively sensitive toexchange rate movements. Owing to an excess of foreign currency assets overforeign currency liabilities, they are exposed to a risk from stronger appreciation ofthe koruna (see Chart IV.32).

Sensitivity analysis: credit risk and the role of real estate pricesFrom the point of view of financial institutions, residential and commercial real estateis not only an investment instrument, but also collateral, especially for loans tohouseholds and corporations. The buoyant growth in loans for house purchase andloans to developers, often secured by real estate, raises the question of what the effecton banks' credit risk would be if an increase in unpaid loans was accompanied by adecline in the prices of real estate used as collateral in the event of a mass sell-off.

The stress tests assume that new default loans have a loan-to-value (LTV) ratio of100%.86 This is quite a radical assumption, as the average values are about 45% for


86 If the standardised approach is applied, only collateral fulfilling the condition of over-securing is acceptable according to the prudential rules (Basel II and a CNB decree define the relationship between the value of real estate and the exposure such that the value of real estate must significantly exceedthe exposure and fully cover the principal, interest and fees). The exposure can be divided into a securedpart and an unsecured part, but no specific LTV is defined for the secured part. The current rise in realestate prices in the Czech Republic, and hence the increasing collateral value, reduces the risk of banksincurring losses in the event of voluntary sales of pledged real estate by debtors.

Baseline scenario SOLVEPost-test SOLVE Required/available solvency margin

Scenario A SOLVEBefore allocation of profit and eq. provisions Post-test SOLVE Required/available solvency margin

Scenario B SOLVEBefore allocation of profit and eq. provisions Post-test SOLVE Required/available solvency margin

Scenario C SOLVE Before allocation of profit and eq. provisions Post-test SOLVE Required/available solvency margin

315 301 327286 290 28335 34 35

315 301 327249 286 220284 288 28035 35 36

315 301 327268 326 222285 289 28135 35 36

315 301 327200 125 260250 217 27640 46 36

TABLE IV.12Solvency and insurance company test results(%)

Total Life Non-lifeInsurance type

Source: CNB

Overall impact of shocks (p.p.)Interest rate shockExchange rate shockCredit shockEquity shockProperty price shock

0.5 -5.8 0.5 -16.71.1 -0.2 4.6 -20.3

-1.0 -2.3 -0.1 7.20.0 0.0 0.0 0.00.0 -3.3 -3.2 -3.40.3 0.0 -0.7 -0.1

TABLE IV.13Impact of shocks in pension fund stress tests(capital adequacy; p.p.)


Note: Calculation for June 2007, derived for illustration from bank capital adequacymethodology in 2006.

Scenario type

CHART IV.32Potential impact of exchange rate shock from open exposure(% of 2007 current profit)

-40-20

020406080

100


Banks Insurers Pension funds

Source: CNBNote: CZK appreciation (Baseline, A, B) and CZK depreciation (C); impact before potentialhedging with currency instruments.

200520062007

43 17 49 8143 27 53 7445 38 55 65

TABLE IV.14Ratio of mortgage loan value to property (collateral)value(LTV in %)

Weighted av. Lower decile Median Upper decileYear

Source: CNBNote: Banks including building societies.


both loans for house purchase and loans to corporations secured by real estate (seeTable IV.14). Moreover, the continuing property price growth is further decreasingthe LTV ratio for existing loans. On the other hand, it can be assumed that entitiesthat obtained loans equalling the value of the collateral do not have sufficientreserves and also face higher interest rates and are more prone to default owing tothe riskier nature of such loans.

A simple sensitivity analysis of loans for house purchase assumes that an increasein the share of loans in default would be accompanied by a fall in real estate pricesof the same extent.87 Problem debtors or banks themselves would sell the collateralon the real estate market, which, in the event of high volumes, would causeproperty prices to decline. This simple test demonstrated the banking sector's highresilience to a mortgage loan portfolio shock given the above-mentioned radicalassumptions. Taking into account aggregate loan developments, the banking sectorshould withstand an increase in the share of default mortgage loans in totalmortgage loans of up to 25% if the return on the voluntary sale of the pledged realestate was at least 75% of its value (see Chart IV.33).

In the model-consistent stress tests, a property price decline can be incorporatedinto the calculation of the effects on capital. In the aggregate calculation, thisrepresents a partial indirect effect of credit risk. Scenario B − "developers in crisis"− assumed a relatively pronounced shock on real estate prices, so its effects are thestrongest. Although the scenario assumed a one-off slump in prices of 30%, theimpact on credit risk and capital adequacy was very low (see Chart IV.34).

Sensitivity analysis: the role of interest rate riskAs regards scenario C ("loss of confidence"), a sensitivity analysis of the impact ofa smooth increase in the interest rate on banks, insurance companies and pensionfunds was performed. The tests assume a shift at the long end of the yield curve ofone-half, i.e. the increase in long-term interest rates (five years or more) is one-halfof the increase in short-term interest rates. The sensitivity analysis showed that thecapital adequacy of the banking sector would fall below the regulatory minimum ifshort-term interest rates rose by more than 4.4 percentage points (see Chart IV.35).By contrast, pension funds are much more sensitive. If we assume a hypotheticalcapital adequacy ratio of 8% for pension funds, their capital would fall to zero ifinterest rates picked up by around 4 percentage points and no additional capitalinjection was provided.

The high sensitivity of the bond portfolios of insurance companies and pensionfunds to interest rate changes is due to their term structure. Whereas in lifeinsurance the investments are usually in long-term instruments, the term structurein non-life insurance is determined by the different nature of creation of insuranceprovisions and claim settlement. A unit shock would manifest itself chiefly in a fallin the value of medium- and long-term bonds, most of all in the life insuranceportfolio and also for pension funds (see Chart IV.36).

61

87 For example, a rise in the volume of default loans of 20% would result in a property price decline of20% according to this very simplified assumption.

CHART IV.33Simple test for mortgage loans(%; 2007)

0

4

8

12

0 10% 20% 30% 40%

Growth of default mortgages and decline in real estate prices

Pre-shock CARCAR after credit shock CAR after credit shock and sale of collateral at a lossCAR after credit shock and sale of collateral(incl. profit allocation)

Source: CNBNote: Scenarios of additional defaults as 10−40% of mortgage loans becoming defaultloans. Banks or clients would sell collateral at 90−60% value.

CHART IV.34Impact of change in property prices on credit risk(capital adequacy; p.p.; 2007)

-0.15

-0.10

-0.050.00

0.05

0.10

Baseline Scenario B Scenario C

Loans to households Loans to corporations

Source: CNBNote: Scenario A does not assume any change in property prices..

CHART IV.35Smooth change in interest rate for selected scenario(capital adequacy; %; 2007)

-5

0

5

10

15

0% 0.5% 1% 1.5% 2% 2.5% 3% 3.5% 4% 4.5%

Banks Insurers Pension funds

Source: CNBNote: Calculation for scenario C.

CHART IV.36Sensitivity of bond portfolio to change in interest rates(by maturity basket; CZK billions; %; 2007)

0

2040

60

80

90 days 91-180days

181-360days

1-2years

2-5years

> 5years

-5.0-4.0-3.0-2.0-1.00.0

LI - real value of bond portfolio (CZK billions)NLI - real value of bond portfolio (CZK billions)PF - real value of bond portfolio (CZK billions)LI - impact of 1% interest rate growth (% of portfolio)NLI - impact of 1% interest rate growth (% of portfolio)PF - impact of 1% interest rate growth (% of portfolio)

Source: CNBNote: Impact of 1% interest rate growth derived from net bond portfolio value (right-handscale). Residual maturity of portfolio derived from modified duration. LI - life insurance, NLI - non-life insurance, PF - pension funds.

Box 7: Stress testing of banks' balance-sheet liquidity

The turbulence afflicting world financial markets since the summer of 2007underlined the key significance of balance-sheet liquidity for ensuring the


efficient operation of the banking sector.88 The drying-up of market liquidityon many financial markets caused problems with financing of the balancesheet and off-balance sheet activities of several banks. This promptedregulatory and supervisory authorities to discuss balance-sheet liquiditymanagement policies with the supervised entities, with the aim of reducingthe risk of such a situation recurring in the future. Stress tests, which to someextent can help banks to assess the potential negative impact of unlikelystress scenarios on their liquidity positions and subsequently set a protectivecushion against potential liquidity risk, are one of the sub-instruments ofliquidity risk management.

Simple stress testing of banks' balance-sheet liquidity is also the focus of thisbox. Only an illustrative stress test was applied, involving a few verysimplifying assumptions (no distinction is made between the business modelsof the individual banks, different maturity types are used for individual assetsand liabilities, the test uses only historical data and does not account for theresponse by the central bank and other implications for the financialmarkets).89 This test serves as an introduction to the methodology of balance-sheet liquidity stress testing, which has not previously been carried out in theanalyses of the Czech financial sector's stability. The test was appliedseparately to 37 banks active on the Czech market. It attempted to answerthe question of whether each of the banks is able to withstand a markedoutflow of its balance-sheet liquidity for 21 days,90 assuming that liquiditycannot be obtained externally (from the central bank, another bank oranother sector). Liquidity outflow is expressed as a loss of clients' confidencein the bank, as reflected in panic deposit withdrawals (a bank run, see Table IV.1 Box). Gap analysis, based on expected cash flows comparing theexpected inflows from the liquidation of selected assets and expectedoutflows from deposit withdrawals, was used to quantify balance-sheetliquidity. The test examines whether the inflow of liquidity for each bankequals or exceeds its potential outflow, i.e. whether a non-negative gap ismaintained.

Three basic liquidity indicators were chosen for the assessment of balance-sheet liquidity (see Table IV.2 Box).91 The starting scenario involves an outflowof each bank's liquidity, as reflected in panic deposit withdrawals currentdeposits. This scenario has two variants, a strong one and a weak one,consisting, respectively, in withdrawals of 5% and 2% of existing demanddeposits per day and 1% and 0.5% of existing time deposits per day. Thedaily percentage of withdrawals was set according to past domestic andforeign experience (for example, it was 3.5% in the case of IPB shortly beforethe imposition of conservatorship and around 5% in the case of NorthernRock in the UK). The scenario is supplemented with other risks taking the

88 A bank's balance-sheet liquidity expresses its ability to meet its obligations in a corresponding volumeand term structure.

89 The assessment of balance-sheet liquidity was applied only to three basic liquidity indicators and thetest uses a simplified model of a bank's balance sheet. The assumptions do not include interbank contagion or lack of confidence in the banking system as a whole.

90 The maximum time for monitoring of the effects of the stress situation on the banks was set at 10 and21 days by expert judgement. It is usually chosen according to the type of scenario selected (e.g. a weekin the case of a failure of a major payment and settlement system, two months to simulate a deep andsudden crisis, etc.).

91 Instead of the "liquid assets" indicator, "assets with a maturity of up to 7 or 30 days" were also usedin the tests. As the results were similar, they are not presented in the box.

Withdrawal of demand deposits per dayWithdrawal of time deposits per dayLiquid assets: available per dayOther assets: available per dayFall in prices of liquid assets (govt bonds)Fall in prices of other assets

5 21 0.5

95 951 1

0.5 0.50.5 0.5

TABLE IV.1 (BOX)Scenario characteristics and shock sizes in stress test(%)

VariantStrong WeakBank run

Source: CNB

LA/Ai)

LA/FLii)

LA/DDiii)

24.00 2.44 12.39 3.15 22.72 2.3926.35 1.85 13.98 3.08 25.21 2.2246.27 2.31 17.18 3.34 29.99 2.83

TABLE IV.2 (BOX)Summary of stress test results after ten days of stress(values express situation of average bank)

Pre-shock Post-shock

RatingValueRatingValueStrong variant Weak variant

Rating v)Value iv)

Selectedindicators

Source: CNBNote: i) liquid assets to total assets,ii) liquid assets to total financial liabilities, iii) liquid assets to total demand deposits, iv) value given in %, v) assessment of size of indicator (1 = low risk, 4 = high risk),

rating thresholds set according to international values.

CHART IV.1 (BOX)Resilience of individual banks to outflow of balancesheet liquidity(x-axis: number of days bank could withstand outflow of balance sheet liquidity; y-axis: number of banks)

05

10

15202530

3540

0. 2. 4. 6. 8. 10. 12. 14. 16. 18. 20. 21.

Strong variant Weak variant

Source: CNB


An alternative indicator of banking sector stabilityThe 2006 Financial Stability Report featured a thematic article introducing analternative indicator of banking sector stability − the banking sector stability index.This index was constructed as a weighted average of sub-indicators of the financialsoundness of the banking system and includes the standard areas used in manyfinancial soundness indicators (e.g. the IMF indicators). The sub-indicators of capitaladequacy, asset quality, profitability, balance-sheet liquidity, currency risk and creditrisk are constructed using financial ratios and normalised so as to express thenumber of standard deviations from the historical average. All partial indicatorswere converted such that an increase means an improvement and a decreasemeans a deterioration.

The aggregate banking stability index continued to decline slightly in 2007 (seeChart IV.39). This was due chiefly to a continuing decline in balance-sheet liquidityand an increase in interest rate risk amid broadly unchanged profitability, capitaladequacy and asset quality (see Charts IV.37 and IV.38). As in the previous year, thisresult can be interpreted as a consequence of credit expansion in a situation ofrelatively low interest rates. Credit expansion causes balance-sheet liquidity todecline and intensifies the time mismatch between assets and liabilities, therebycontributing to a rise in interest rate risk. Asset quality, measured by the share ofdefault loans, does not deteriorate despite the credit expansion, as the rise indefault loans is diluted by the growth in new loans.

Overall, the evolution of the index can be viewed as a reflection of the optimisationof capital and its return towards the historical average in a situation of a creditboom rather than as a decline in the banking sector's resilience to shocks. This is confirmed by comparison with the results of the standardised stress test (seeChart IV.39). Capital is sufficient to cover the losses given the current quality andmanagement of assets. However, the process of optimising capital utilisation mayraise certain questions regarding future developments. A cooling of the economywould lead to a rise in default loans, which would affect profitability. Capitaloptimisation would be replaced by a slowdown in credit growth and possibly aforced increase in capital to maintain capital adequacy, as indicated by the resultsof the stress tests of the alternative scenarios.

63

92 Generally speaking, a bank attempts to balance the inflows and outflows of liquidity at different assetbasket maturities, at the same time holding sufficiently liquid assets in case of an unexpected mismatchin the balance sheet. If such a mismatch occurred and the bank tried to remove it by means of a largesell-off of assets, it could result in a decline in the prices of those assets given insufficient elasticity ofdemand.

CHART IV.37Sub-indicators of banking sector stability(standard deviations from historical average)

-3

-2

-1

0

1

2

3

97 98 99 00 01 02 03 04 05 06 07 08

Capital adequacyAsset qualityProfitability

Source: CNB

CHART IV.38Sub-indicators of banking sector stability (standard deviations from historical average)

-3

-2

-1

0

1

2

3

97 98 99 00 01 02 03 04 05 06 07 08

Liquidity Interestrate risk

Exchangerate risk

Source: CNB

CHART IV.39Banking sector stability index(standard deviations from historical average; capital adequacy in %)

-2

-1

0

1

2

97 98 99 00 01 02 03 04 05 06 07 080246810121416

Banking stability index (left-hand scale)Capital adequacy (right-hand scale)Post-test capital adequacy (right-hand scale)

Source: CNBNote: Stress test based on historical scenario I (see FSR 2006), combining a rise in defaultloans (of 30%), depreciation of the koruna (by 15%) and an increase in interest rates (by 1 p.p.).

form of a decline in market liquidity and a fall in asset prices.92 A decrease inthe value of liquid assets (government bonds) and non-liquid assets of 0.5%every day is assumed (see Table IV.1 Box).

The results of the test after ten days of stress (see Table IV.2 Box) show aconsiderable decline in the selected indicators for the average bank under the"strong" variant. The "weak" variant did not lead to a significant decrease inthe average value of the monitored indicators. The chart (see Chart IV.1 Box)provides an answer to the question above − whether the bank is able towithstand an outflow of balance-sheet liquidity for 21 days. It shows that 18monitored banks would be able to withstand a strong outflow of liquidity for21 days despite the very strong assumption underlying the test. As regardsthe "weak" variant, only three banks would not be able to withstand theoutflow of balance-sheet liquidity.

PART II − THEMATIC ARTICLES

THE ROLE OF RATINGS IN FINANCIAL SECTOR STABILITY ASSESSMENT 65

THE ROLE OF RATINGS IN FINANCIAL SECTOR STABILITY ASSESSMENT

Petra Davidová and Renata Opravilová, CNB

This article discusses the ratings of financial market participants as one of the indicators of financial stability. Particularattention is paid to the sector of banks and insurance corporations in the Czech Republic. Besides discussing theratings of individual financial institutions, the article explores aggregate ratings. For the banking sector, Fitch's BankSystemic Risk Matrix is used as an aggregate rating for assessing a country's financial stability. For the insurancesector, the article presents for the first time the results of a newly constructed aggregate rating for insurers in theCzech Republic, which performs an analogous role to the matrix mentioned above. The analyses conducted in thearticle confirm that the banking and insurance sectors are showing satisfactory and steadily improving performancefrom the ratings perspective, in line with the conclusions of previous analyses of a different type. An analysis of thecorrelations of rating types and of the correlations between the ratings of banks and those of their ownersdemonstrated that the individual types of ratings of the various agencies are not fully interchangeable and that thereis a positive correlation between the average long-term ratings of banks and those of their owners.

1. INTRODUCTION

Ratings are one of the indicators of the financial soundness of financial institutions and thus also of the overallfinancial stability of the system. They stand alongside other composite indicators which, in aggregate, i.e. by meansof a single score, express the degree of resilience to risks undertaken. In the case of banks and insurancecorporations, these other indicators include capital adequacy ratios, solvency ratios and financial soundnessindicators.93 The advantage of ratings − unlike solvency ratios and financial soundness indicators − is that they takeinto account qualitative information on companies, including expert expectations. On the other hand, it isimportant to emphasise that ratings provide only an estimate of the probability of default, and deducing the levelof other types of risks from this indicator can lead to overestimation of the indicator. The recent subprime mortgagecrisis in the USA shows that a correct understanding of ratings by investors is crucial for financial stability.

In this article we look at bank ratings (section 2) and assess the ability of such ratings to reflect potential supportfrom strategic owners and the state. In this context, we examine the dependence of bank ratings on the ratingsof their owners. Fitch's Bank Systemic Risk Matrix is used to assess the banking sector as a whole. In section 3,we deal analogously with ratings on the insurance market. We present the first-ever construction of an aggregateinsurance sector rating, which to some extent addresses the fact that there are fewer official ratings in theinsurance sector than in the banking sector and that there is no systemic rating for the insurance sector.

Our analysis of banks and insurance corporations is based on the ratings issued by the best-known agencies:Moody's, Standard&Poor's (S&P), Fitch and A.M. Best.94 The methodologies, procedures and symbols used bythese agencies differ, even when same types of institutions and instruments are assessed. The range of entitiesassessed by means of external ratings is enormous.95 In this article we focus on the ratings of banks and insurancecorporations in the Czech Republic. Other areas, such as sovereign ratings (indicating the state's ability to meetthe obligations arising from the loans it accepts or the bonds it issues), corporate liability ratings and the use ofratings as a regulatory tool under Basel II are dealt with in other CNB publications and in the economic literature.96

The role of ratings in the mortgage market crisis has been significant, as stated in section 2.1 of this report.

2. BANK RATINGS

In this section we look at bank ratings. The specific features of bank ratings are described and an analysis isconducted of the ratings of banks and their parents and affiliates. The Fitch matrix is used to assess the systemicbanking risk of the Czech Republic.

93 See Geršl and Heřmánek (2007).94 A.M. Best specialises primarily in the insurance market.95 For example, Moody's alone issues around 40 types of rating depending on the entities assessed.96 See Liška and Vinš (2005), Sůvová, Kozelková, Zeman and Bauerová (2005) and Derviz and Podpiera (2004).

66 THE ROLE OF RATINGS IN FINANCIAL SECTOR STABILITY ASSESSMENT

Specific features of bank ratingsThe banking sector can be characterised by a relatively small number of defaults97 compared, for example, to thecorporate sector. An important role is played here by the fact that banks are subject to state regulation andsupervision and enjoy potential support from the state should they run into difficulties. These support mechanismsare reflected in banks' ratings.

Rating agencies usually issue banks with a rating expressing the unsupported (self-reliant) risk of failure of thebank and a rating evaluating the state's or owner's ability and propensity to support it. These components arecontained, for example, in Fitch's individual and support ratings and joint probability default analyses. The supportrating98 sets a floor for the bank's long-term rating.99 The setting of such a floor significantly reduces the volatilityof long-term ratings when severe adverse factors hit the economy. Likewise, Moody's uses a bank financialstrength rating to assess a bank's own credit risk. In 2007, the agency introduced joint default analyses, whichexplicitly include potential support in banks' final ratings. This leads generally to an increase in long-term ratingsand a slight decrease in bank financial strength ratings.

Table 1 shows the correlations100 between selected types of ratings issued to banks in the CEC5 region101 by thethree main agencies. In all cases the correlations are positive. The most closely correlated are Moody's financialstrength ratings and Fitch's individual ratings; this is consistent with their similar content. The analysisdemonstrates that the individual types of ratings from the various agencies are not fully interchangeable.102 Forthis reason, most large institutions on the market pay for ratings commissioned from all three agencies. Banks inthe Czech Republic likewise have ratings from several different agencies.

97 Fitch (2007a) defines bank default and bank failure as follows: "A bank has defaulted if it fails to make a timely payment of principal and/orinterest. … A bank has failed if it is kept going only by state support or support from a (deposit) insurance fund … if it is kept going only bybeing acquired by some other corporate entity … if it is kept going only by an injection of new funds from its shareholders or equivalent …[or] if it has defaulted". Fitch (2007a) states that over the past 25 years there have been no instances of default among banks in developedeconomies with high support ratings (1 or 2). Up to 2003, there were only seven bank defaults in developed economies (mostly small US institutions); these banks defaulted because they failed to get third-party support.

98 The support rating combines two factors: the state's or owner's ability to support (as expressed by its own long-term rating) and its readinessto support (as estimated by the agency). It ranges from 1 (an implied propensity to support of at least 99%) to 5 (a 40% propensity to support).

99 For example, a bank with a support rating of 1 has a long-term rating floor of A-. In an "AAA" economy, the implicit risk that the "AAA" sovereign will not be willing to support is equal to 44 b.p. (for more details on the calculation, see Fitch, 2007a).

100 For the purposes of these correlations, the qualitative rating scale was converted into a numerical one assuming that the difference betweenthe individual rating scores is the same. However, the rating agencies do not set target quantitative default rates for the individual rating scores;only the historical rates are known. Hence, the rating cannot be precisely quantified and the approach used is something of a simplification.

101 All banks in the Central and Eastern European countries (the Czech Republic, Hungary, Poland, Slovakia and Slovenia) which have a rating.The close correlation results for S&P's ratings are due to the fact that the BankScope database contains very few banks that have ratings fromboth S&P and another agency. The sample contains around 350 credit institutions.

102 In addition to the dependence on the specific sample assessed, we should emphasise in particular:● the diversity of the agencies' models, ● the expert component of the rating, which further increases the uniqueness of each agency's assessment, ● the possibility of a clearer assessment of a given company, in particular as regards the number of unknowns entering futureexpectations, ● the speed at which changes are incorporated into an updated rating, ● the rating scale, as a more detailed scale allows morenuances to be differentiated, ● better knowledge of a specific environment by a particular agency, ● and sometimes also the declared owninterests of the agency.

Table 1 − Correlation matrix for individual types of ratings from Fitch, Moody's and S&P(correlation coefficients; CEC5 region; as of 31 December 2007)

Fitch Fitch Fitch Moody's Moody's S&P support long-term individual long-term financial long-term

rating rating rating rating strength rating rating

Fitch support ratingFitch long-term ratingFitch individual ratingMoody's long-term ratingMoody's financial strength rating S&P long-term rating

1_ 1_ _ 1

0.75 0.566 0.46 10.112 0.23 0.797 0.5 1

1 1 1 0.707 1 1

Source: BankScopeNote: The selected sample of all banks in the CEC5 countries contains very few banks that have a rating both from S&P and from another agency, which meansthat the correlation results are affected.


Bank ratings in the Czech RepublicThe profiles of the external ratings issued by the main international rating agencies between 2000 and 2007illustrate the positive development of the Czech banking sector and the positive macroeconomic trend103 (Table 5and Chart 2). The main reason for the rising ratings are higher profitability and cost-effectiveness of banks,increased competition, a significant increase in service quality, and improved credit risk management (e.g. CNB,2007a). In recent years, banks have been strengthened by the completion of their restructuring processes and bygrowing interest in loans.

The average Moody's financial strength rating for Czech banks is C with a positive outlook, and the average long-term rating in the Czech currency is Aa3 with a stable outlook. Compared to the other CEC5 banking sectors, theCzech system has the highest financial strength rating (at the same level as Austria) and the second-best long-term rating (behind Slovenia) (Chart 1). Like their parents, the largest Czech banks have the highest possiblesupport rating (i.e. an implied propensity to support of 99%).

Chart 2 compares the evolution of the average rating of the three largest Czech banks and that of their owners.The slightly higher rate of improvement of the Czech banks' rating in 2001−2003 is linked with higher relativeprofitability of subsidiary banks in this period (CNB, 2005).

103 The analyses focused on the three largest Czech banks, ČS, ČSOB and KB, which account for around 60% of the assets of banks based in the Czech Republic. This makes the sample sufficiently representative (Derviz and Podpiera, 2004). Besides these banks, ČEB and J&T haveexternal ratings, but their market shares are very small. In the past, Živnostenská banka also had a rating, but it went out of business in 2007.This bank was not included in the analysed sample, as the options for investigating the relationships were limited owing to changes in ownership.

Table 2 − Moody's, Fitch and S&P rating scales

Moody's Fitch Standard&Poor'sLong-term Short-term Fin. strength Long-term Support Individual Long-term Short-term

Excellent

Good

Reasonable

Speculative

Highly speculative

Aaa P-1 A AAA 1 A AAA A-1+Aa1 P-1 A- AA+ 1 A AA+ A-1+Aa2 P-1 B+ AA 1 A AA A-1+Aa3 P-1 B AA- 1 A/B AA- A-1+A1 P-1 B- A+ 1 B A+ A-1A2 P-1 C+ A 1 B A A-1A3 P-2 C A- 1 B/C A- A-2Baa1 P-2 C- BBB+ 2 C BBB+ A-2Baa2 P-2 C- BBB 2 C BBB A-2Baa3 P-3 D+ BBB- 2 C/D BBB- A-3Ba1 Not prime D+ BB 3 D BB+ BBa2 Not prime D BB 3 D BB BBa3 Not prime D- BB 3 D/E BB- BB1 Not prime E+ B 4 E B+ CB2 Not prime E+ B 4 E B CB3 Not prime E+ B 5 E B- CCaa Not prime E C 5 E CCC C

Source: Moody's, Fitch, Standard&Poor's


The evolution of the ratings of Czech banks since 2000 shown in Table 3 indicates that the ratings of Czech banksand their parent banks from S&P correlate most closely. This result can be explained by the different constructionof the ratings of this agency. Unlike Moody's and Fitch, S&P does not have products specifically distinguishingbetween ratings with potential state or owner support and those without it, and probably takes such support intoaccount more in its long-term ratings. The low correlation coefficient for financial strength ratings (Moody's) isconsistent with the logic of this rating, which measures the institution's internal strength regardless of theprobability of owner or state support. The negative correlation of Fitch's ratings in the case of SocGen/KB is dueto the temporary lowering of Société Générale's rating and the steady rise in KB's rating in 2003−2005, whichcorresponds to the higher relative profitability of the subsidiary bank in this period (CNB, 2005).

An aggregate rating for the Czech banking sectorA Fitch rating − the Bank Systemic Risk Matrix (see Fitch, 2007b) − can be used to assess the financial stability ofa country and the stability of its financial system as a whole. The matrix combines two indicators: a BankingSystem Indicator (BSI) and a Macro-Prudential Indicator (MPI). The BSI measures the true "intrinsic" bankingsystem strength regardless of potential state or owner support and is calculated as an asset-weighted average ofindividual bank ratings. The BSI ranges from A (very high quality) to E (very low quality). The MPI tracks three keyindicators: the ratio of private sector credit to GDP; the real effective exchange rate; and equity prices. Hence, theMPI takes into consideration the existence and severity of a set of macroeconomic circumstances that in the pasthas anticipated full-blown systemic crises in the banking sector. It ranges between 1 (low vulnerability of thesystem) and 3 (high vulnerability). A score of 3 applies when the ratio of private sector credit to GDP is more than5% and either equity prices are more than 40% above trend or the real effective exchange rate is more than 9%above trend. The advantage of this approach is that it takes account of the fact that weak spots in the banking

Fran

ce

Belg

ium

Italy

Ger

man

y

Aus

tria

Cze

ch R

ep.

Hun

gary

Slov

enia

Pola

nd

Slov

akia

Average financial strength rating (left-hand scale)Average long-term rating of foreign-currencybank deposits (right-hand scale)

Aa1

Aa2

C-

C+

B-

D+

Aa3

A1

2000 2001 2002 2003 2004 2005 2006 2007

Average rating of three largest Czech banks,average for all three agencies

Average rating of their parent banks,average for all three agencies

A+

A-

A

Chart 1 − Comparison of average ratings of CEC5 countriesand home countries of parents of Czech banks and branchesoperating in Czech Republic(as of 31 December 2007; asset-weighted average)

Chart 2 − Average ratings of three largest Czech banks andtheir parent banks(long-term foreign currency ratings, Moody's, Fitch and S&P; simple average)

Source: Moody's Source: BankScope, CNB calculation

Table 3 − Correlation of series of ratings of largest Czech banks and their parent banks(correlation coefficient; 2000−2007; yearly observations)

Moody's financial Moody's Fitch S&P strength rating long-term rating long-term rating long-term rating

SocGen/KBKBC/ČSOBErste/ČS

Owner/bank

0.25 0.41 -0.16 0.65-0.27 0.22 0.60 0.650.13 0.76 0.47 0.64

Source: BankScope, CNB calculation


system (always reflected in banks' individual ratings and aggregated in the BSI) can have systemic impacts in anadverse macroeconomic environment (reflected in the MPI). We can see some limitations in this matrix in that thethree-point MPI scale makes it impossible to differentiate between countries where average annual credit growthis around 20% (the case of the Czech Republic) and countries where it is around 40% (e.g. Latvia and Estonia) −both these groups get MPI = 2. The result is that half of the 87 countries that Fitch includes in this matrix are incategory MPI = 2 and 9% are in category MPI = 3104 − i.e. roughly 60% of the countries lie in the higher-risk partof the matrix. Table 4 shows the results for the EU member states. It is clear from the table that the Czech Republichas the best bank systemic risk rating of all the CEC5 countries.105 This matrix seems to be an appropriatecomplement to the IMF's Financial Soundness Indicators as discussed in Geršl and Heřmánek (2007).

The results of the analysis show that the improving ratings of Czech banks and the clear convergence trendtowards their owners' ratings indicate a steady rise in the performance and stability of the Czech banking system(e.g. CNB, 2007b). At the same time, the broadly similar ratings of their affiliates generate no concerns aboutpotential problems arising and spilling over into the domestic financial system. An analysis of the correlationsbetween rating types and of the correlations between the ratings of banks and those of their owners

104 These are non-European countries, e.g. Canada, Korea, South Africa, Iran and Azerbaijan.105 The different MPIs for Hungary and Poland can be explained by the slower credit growth in these countries in 2005 and 2006, which did not

exceed 15% − the limit for a switch to MPI = 2 (Fitch, 2007b).106 The rating scales are given in Table 2.

For more details on definitions and rating scales, see www.moodys.com, www.fitchratings.com and www.standardandpoors.com.

Table 5 − Comparison of ratings of Czech banks and their parent banks,branches and affiliates in CEC5 countries(ratings of main rating agencies;106

as of 31 December 2007)

Moody's S&P Fitch

Three largest Czech banks

Parent banks

Branches

Affiliates ofthree largestCzech banks

KBČSOBČSSocieté Generale (France)KBC Bank (Belgium)Erste (Austria)ING Bank N.N.HSBC Bank plcCommerzbank AGDeutsche Bank AGCalyon S.A.Fortis Bank SA/NVSKB Bank d.d. (Soc. Gen. Slovenia)Slovenska sporitel'na(Erste Slovakia)Erste Bank HungaryKB (Soc. Gen. Slovakia)Nova Ljubljanska Banka(KBC Slovenia)K&H Bank Rt. (KBC Hungary)Kredyt Bank S.A. (KBC Poland)

A1/P-1/C A+/A-1 AA-/1/B/CA1/P-1/C n.a. A+/1/B/CA1/P-1/C A/A-1 A/1/B/CA1/P-1/B AA/A-1+ AA/1/A/B

Aa2/P-1/B- AA-/A-1+ AA/1/A/BAa3/P-1/C A/A-1 A/1/B/CAa1/P-1/B AA-/A-1+ AA/1/A/B

Aa3/P-1/n.a. AA-/A-1+ AA/1/BAa3/P-1/C+ A/A-1 A/A-/n.a.

Aa1/P-1/B AA/A-1+ AA-/1/BAa1/P-1/C AA-/A-1+ AA/1/B/C

Aa2/P-1/C- AA-/A-1+ AA-/1/n.a.

A1/P-1/D+ n.a. n.a.

A1/P-1/C- A-/n.a. A/1/C/DA2/P-1/C+ n.a. n.a./1/n.a.

n.a. n.a. n.a.

Aa3/P-1/C n.a. A-/1/C

A2/P-1/C BBB/n.a. A+/1/D

A2/P-1/D BBB/n.a. A+/1/D

Table 4 − Bank Systemic Risk Matrix for EUmember states(Fitch, 2007; MPI = Macro-Prudential Indicator; BSI =Banking System Indicator)

MPI

BSI 1 2 3

A LuxembourgNetherlands

SpainSwitzerland

United KingdomB Austria Belgium

Germany Czech Republic

DenmarkEstoniaFinlandFranceGreeceIreland

PortugalItaly

SwedenC Cyprus Latvia

MaltaSlovakiaSlovenia

D Hungary BulgariaPoland Lithuania

RomaniaE

Source: FitchNote: MPI = 1 low vulnerability; MPI = 3 high vulnerability

Source: Moody's, S&P, Fitch, BankScopeNote: Moody's (long-term rating/short-term rating rating/financial strength rating), S&P (long-termrating/short-term rating), Fitch (long-term rating/support rating/individual rating).

demonstrated that the individual types of ratings from the various agencies are not fully interchangeable and thatthere is a positive correlation between the average long-term ratings of banks and those of their owners.

3. RATINGS ON THE INSURANCE MARKET

This section discusses the individual ratings of institutions operating in the insurance market. The small numberof institutions with ratings107 is a limiting factor for the assessment of this sector in the Czech Republic. We triedto solve this problem by constructing our own aggregate rating of the insurance company sector.

Specific features of insurance company ratingsThe agencies issue insurance companies with two main types of ratings. The financial strength rating gives anindependent expert opinion on an insurer's ability to pay under its insurance policies and contracts in accordancewith their terms (Standard&Poor's, 2007). Its ability to meet other debt obligations, especially those associatedwith issues of fixed-interest securities, is the subject of a debt rating. The financial strength rating and the debtrating can differ from each other, as they serve different purposes and do not measure the same risks.

Only some of the insurance companies with one rating request a rating from another agency. Most globalinsurance market participants opt for a rating from A.M. Best. The literature states that the ratings issued byMoody's and S&P tend to be lower than those issued by A.M. Best (e.g. Pottier and Sommer, 1999).108 We notedthe same fact in our sample of foreign insurers and reinsurers relevant to the Czech insurance market, with A.M. Best assigning a higher rating than the other agencies in 13 cases out of 18.

Unlike banks, which can obtain support from their owners or the state if they run into difficulties, insurance companiescan obtain extraordinary funds to meet their obligations from contractual reinsurers as well. These institutions play animportant role in the financial stability of the insurance sector, as reinsurance represents a transfer of risk from insurerto reinsurer.109 When a rating is assigned to an insurer, the financial strength of the relevant reinsurers is taken intoaccount as well. A high-quality reinsurer should mean a higher rating for the reinsured insurer. Rating agencies stateexplicitly that they take the financial strength expressed by the relevant ratings of contractual reinsurers intoconsideration in the rating process (A.M. Best, 1996). Contracts between insurers and reinsurers often include ratingclauses containing a trigger mechanism that comes into operation when the reinsurer's rating falls below a certain level(ECB, 2006).110 These links between the financial strength ratings of insurers and those of their contractual reinsurerswere used to construct an aggregate rating for the Czech insurance sector, which is presented later in this section.

Insurance company ratings in the Czech RepublicČeská pojišťovna is currently the sole domestic insurance company that has a rating from two internationallyrecognised agencies. ČSOB Pojišťovna is the only other insurer with a rating from one of these agencies. TheCzech insurance market thus lags behind the advanced economies in terms of the number of companies withratings.111 The two insurers mentioned above have sizeable market shares.112 Their ratings are of investment gradeand are steadily rising113 in line with the conclusions of CNB documents (CNB, 2006, 2007a) concerning the qualityand stability of the insurance sector.


107 Only two out of the total of 34 insurance companies based in the Czech Republic (i.e. 6%) currently have a rating from an external agency. 108 A.M. Best started issuing ratings to insurance companies back in 1899 and for a long time specialised exclusively in this sector. Other

agencies entered the insurance segment of the market in the 1980s and the ratings newly issued to insurance companies had to conform totheir previous practices.

109 The transfer of risk to a reinsurer simultaneously increases the insurance capacity of the insurer, which can insure more risks and generate abigger profit. This occurs only to the extent to which the reinsurers are capable of meeting their liabilities.

110 For example, the rating clause may state that if the reinsurer's rating is downgraded below a certain level (the rating trigger), the insurer mayrequire the reimbursement of part of the premiums it paid or terminate the contract.

111 This situation has two apparent causes. Low pressure from other market participants is exacerbated by the price of ratings, which can be highfor relatively small institutions. Subsidiaries and branches of major international insurers also operate in the Czech insurance market. Their ratings can sometimes be used for a simplified assessment of an entity in the Czech Republic.

112 The Czech Insurance Association gives shares of 31% and 7% of total premiums written for Česká pojišťovna and ČSOB Pojišťovna respectively. Both these insurers are members of the group of five large insurance corporations (CNB, 2007b).

113 The agencies find strengths in several areas. In the case of Česká pojišťovna they emphasise its long-term stable business performance, its good capital adequacy and liquidity, and also its strong market position. A major reason for its February rating upgrade from S&P was its incorporation into Generali PPF Holding. ČSOB Pojišťovna's rating is underpinned by its growing strategic importance for the parent KBCInsurance NV, including its increasing integration into the KBC group, and its improving financial performance.


It is clear from Chart 3 and Table 6 that the ratings of the two Czech insurers are lower than those of the ownersof the other large domestic insurers and those of the relevant reinsurers. This is due primarily to the smaller sizeof the domestic institutions and the limiting value of the Czech Republic's sovereign rating.

Construction of an aggregate rating for the Czech insurance sectorIn this section we present our proposed calculation of an aggregate rating for the Czech insurance sector. The aimis to estimate an average rating for insurance companies in the Czech Republic which might, to some extent,make up for the lack of ratings actually assigned by agencies. The resulting value may, in line with the definitionof the insurance company financial strength rating, serve as an indicator of the sector's ability to pay under itsinsurance policies and contracts in accordance with their terms. The result may also be used as one of the inputsfor a comprehensive assessment of the sector.

We are chiefly interested in the category of large insurance companies, which accounts for roughly 75% ofpremiums written in the Czech Republic (CNB, 2007b). Our proposed method takes into account the relevant validratings of insurers, reinsurers and owners, the level of reinsurance of liabilities, and the weights of the individualinsurers. This information enters the calculation via the parameters given in Table 7.

Table 6 − Ratings of the largest reinsurers of insurersactive in the Czech Republic (as of 31 December 2007)

Country S&P Moody's Fitch Best

Münchener RückversicherungSCOR ParisSwiss Re GermanySwiss Re FrankonaRückversicherungNational Indemnity CompanyHannover ReWiener StädtischeTransatlantic Reinsurance, Paris branchNew Reinsurance Company

DE AA- Aa3 AA- A+FR n.a. n.a. n.a. A-DE AA- Aa2 AA- A+

DE AA- Aa2 AA- AUS n.a. Aaa AAA A++DE AA- A3 A+ AAT A+ n.a. n.a. n.a.

FR/US n.a. n.a. n.a. A+CH AA- n.a. n.a. A+

Source: Databases of the rating agencies

Aa1

Baa3

AA

Allianz Generali

ING

Kooperativa

ČSOBPojišťovna

Česká pojišťovna

2007

Česká pojišťovna

2008

Standard & Poor's

Moo

dy's

BBB

Chart 3 − Ratings of the largest insurers active in the CzechRepublic and their owners (as of 31 December 2007)

Source: Databases of the rating agenciesNote: Both ratings have been issued to insurers based in the Czech Republic only inthe case of Česká pojišťovna, whose rating by S&P rose to A in February 2008. ČSOBpojišťovna has been issued with an S&P rating. In all other cases, the ratings pertainto the owners.

Table 7 − Input parameters

Weight of insurer

An insurance company's market share is usually expressed in terms of its share in premiums written. In our case, we opted for a variable that better matches

the character of the financial strength rating, which reflects the ability to meet obligations arising under insurance policies and contracts. The weight is

therefore given by the insurance company's share in total technical provisions, which represent obligations arising under insurance policies and contracts.

Reinsurance

Reinsured amounts are divided into three categories according to the amount pertaining to the particular reinsurer and according to the quality of the

reinsurer as expressed by its rating. Of the total of almost 300 contractual reinsurers of large insurance companies, we chose the 15 largest reinsurers, which

account for 73% of the total reinsurance. The reinsurers and the amounts reinsured by them are divided into:

- the set of reinsurers having a financial strength rating from at least one of the four main agencies (12 reinsurers)

- the set of reinsurers that do not have such a rating (3 reinsurers)

- the remaining 27% of the reinsured amount, entering the calculation in a unified way with a single rating.

Ratings

The insurance companies are progressively assigned up to seven types of ratings (from agencies or calculated), which are derived from the existing ratings

of the insurers and their owners and reinsurers.


The five largest insurers have around 25% of their non-life insurance liabilities reinsured with contractual partners.Including life insurance, which is rarely reinsured, the reinsurance ratios among the individual large insurers fall to 5%−22% of premium liabilities.114 The calculation covers the 15 largest reinsurers out of the total of almost300 reinsurers that have contracts with the five large insurers. These 15 reinsurers account for 73% of the totalreinsurance, while 67% of them have at least one rating and 53% at least two ratings.

Ratings are assigned to the individual insurers according to the priorities shown in Table 8. If an insurer has a rating/ratings from an external agency, this rating − or, where relevant, its average rating − is assigned to it asa whole. In other cases, the ratings of contractual reinsurers are taken into account for reinsured liabilities. Fornon-reinsured liabilities, a reduced rating of the owner is used. The resulting calculated rating of the insurer is acombination of the weighted ratings of the reinsurers and the weighted adjusted ratings of the owners. Theratings of the individual institutions weighted by their shares in technical provisions then enter the calculation ofthe rating of the entire group of large insurance companies.

114 Insurance companies employ various strategies for reinsuring their liabilities. The largest domestic insurers have around 100 such contractualpartners. Other insurers, by contrast, have exclusive partners. Reinsurance is often arranged with a member of the same financial group.

Table 8 − Procedure for assigning ratings to insurers

Insurer based in Czech Republic has rating(s)

Insurer based in Czech Republic has no rating

Reinsured liabilities

Reinsurers (out of 15 largest) with rating(s)

One rating

Two or more ratings

Reinsurers (out of 15 largest)

with no rating

Reinsurers other than 15 largest

Non-reinsured liabilities

Rating assigned to total liabilities

pertaining to reinsurer

Average rating

assigned to total liabilities

pertaining to reinsurer

Average of three ratings*assigned to

liabilities pertaining to these

reinsurers

Average rating of 12

(out of 15 largest)reinsurers assigned

to liabilities

Owner's adjusted rating,

net of risk surcharge, assigned

to liabilities

One rating

Two or more ratings

Average rating

assigned to insurer as a whole

Single rating

assigned to insurer as a whole

Note: The three ratings entering the calculation are: 1. the owner's adjusted rating, net of a risk surcharge; 2. the simple average of the ratings of 12 (out of the15 largest) reinsurers = a constant; 3. the average of the ratings of reinsurers (out of 15 largest) weighted by the contractual amounts for a given insurer.

➡

➡

➡

➡ ➡ ➡ ➡ ➡ ➡ ➡


When setting the risk surcharge for the owner's rating, which our method employs to calculate the rating for thenon-reinsured portion of the liabilities, we used the relationship between the existing ratings of domesticinstitutions and their owners. In the case of insurers, such a relationship exists between ČSOB Pojišťovna and KBCInsurance N.V. of Belgium, whereas in the case of banks such relationships exist between all the Czech-basedbanks analysed. Our method for calculating the rating of insurance companies draws on the conclusions of the analysis of the ratings of Czech banks and their owners described in section 2 of this article and presented inChart 2. We start by assuming that the relationships that apply in the banking sector also exist for insurancecompanies and that the evolution of the ratings of foreign owners can be a good guide for estimating theevolution of the ratings of domestic insurers. That our approach is justified is confirmed by Česká pojišťovna'sFebruary rating upgrade, which demonstrated how important it is for an insurance company in the Czech Republicto have ownership or operational links with a strong foreign partner. In the calculation we lowered the rating bythree grades. The resulting ratings calculated for the group of large insurance companies in the Czech Republicare illustrated in Chart 4.

At the end of 2007, our calculated average rating for insurance companies in the Czech Republic, expressed usingthe scale employed by S&P, stood at BBB+/BBB, i.e. investment grade.115 Taking into account Česká pojišťovna'srating upgrade issued by S&P in February 2008, the overall calculated rating increases by one grade. Thisimprovement corresponds to Česká pojišťovna's weight in the sector and to the rise in its rating by three grades.

The calculated aggregate financial strength rating approximates the quality level of the Czech insurance sector asa whole as regards insurers' ability to meet their obligations arising under insurance policies and contracts andthus to have a positive impact on the financial stability of other economic sectors acting in the role of their clients.

Our constructed indicator has some limitations. In a situation where an insurer's reinsured liabilities are on the risebut the relevant reinsurer's rating is downgraded, the insurer's calculated rating could decrease.116 Another

115 This calculated rating took values of Baa1/Baa2 according to the Moody's scale, BBB+/BBB according to the Fitch scale and B+/B according tothe A.M. Best scale.

116 However, this could only happen if the reinsurer's rating was lower than the reduced rating of the insurer's owner. Looking at the data in Chart 3 and Table 6, this is not very likely.

Českápojišťovna

2007

Česká pojišťovnasince

7 February2008

ČSOBPojišťovna

Largeinsurers 2007

calculation

Largeinsurers2008

calculation

BBB

A-

BBB+/BBB

A

A-/BBB+A

BBB-

BB+

BBB

BBB

Chart 4 − Existing ratings of insurers in Czech Republic and calculated aggregate ratings according to S&P scale (as of 31 December 2007)

Source: S&P database, CNB, CNB calculationNote: The group of large insurers in 2008 reflects a rise in Česká pojišťovna's rating on 7 February 2008 by S&P, while all other ratings keep their end-2007 values. The group of large insurers accounts for around 75% of the Czech market. The bar height represents the average rating attained.


limitation is that the relationship between the ratings of institutions operating in the Czech Republic and theratings of their owners is based on a small number of observations, so the calculation is affected by a small sampleof existing relationships. The fact that the ratings between parent and subsidiary insurers do not necessarily movein the same way, owing to differing performances of specific institutions and differing macroeconomic trends inindividual countries (leading, for example, to a change in the sovereign rating in one relevant economy), can alsobe regarded as something of a limitation.

The method outlined above cannot be applied to the banking sector, as it does not feature institutions withfunctions analogous to those of reinsurance companies. The Czech banking sector, moreover, has a far higherrating coverage than the insurance sector.117 The need to construct an aggregate rating is thus less pressing thanin the case of insurance companies. For the same reason, one can view the construction of an aggregateinsurance sector rating as being less important in national sectors where the majority of insurers have ratings (interms of number and market share) and in countries where the insurance sector represents only a negligible partof the financial sector.

4. CONCLUSIONS

The ratings of Czech banks and their parents and affiliates, as well as the systemic rating of the Czech bankingsector, confirm that these institutions are in sound financial health, including as regards the potential spillover ofproblems within banking groups. Also, the individual ratings and our constructed aggregate rating for Czechinsurers confirm the positive assessment of the insurance sector conducted earlier by means of stress testing. Theanalysis confirmed the mainly positive correlations between the ratings of Czech banks and those of their owners.At the same time, however, the discussion in this article demonstrated that the individual types of ratings or sameratings from the various different agencies are not interchangeable. Rating agencies do not automatically reflectan owner's rating downgrade in its subsidiary institution's rating, as confirmed by a negative correlation betweenthe ratings of banks and those of their owners in some cases and also by current developments − SociétéGénérale's rating was recently downgraded as a result of the liquidity crisis following the subprime crisis in theUSA, but none of its subsidiaries had their ratings lowered.

Some rating products, such as Fitch's Bank Systemic Risk Matrix or our proposed aggregate insurance companyrating, can be used to complement the Financial Soundness Indicators. Both these indicators, however, have somelimitations, just like ratings themselves do. None of the ratings discussed here can be used to assess the financialstability of the system or the stability of individual banks and insurers without knowledge and application of othersupporting indicators and tools.

REFERENCES

A.M. BEST (1996): Best's Key Rating Guide, Property-Casualty Edition

CNB (2005): Financial Stability Report 2004, Czech National Bank


CNB (2007a): Financial Stability Report 2006, Czech National Bank

CNB (2007b): Financial Market Supervision Report 2006, Czech National Bank

DERVIZ, A., PODPIERA, J. (2004): Predicting Bank CAMELS and S&P Ratings: The Case of the Czech Republic,Working Paper No. 1, Czech National Bank

117 The differences are 17 p.p. as regards numbers and 19 p.p. as regards market shares in favour of banks.


ECB (2006): Financial Stability Review, June 2006

ECB (2007): Large Banks and Private Equity − Sponsored Leveraged Buyouts in the EU, ECB, April 2007

FITCH (2002): Bank Support in the Developed World, Fitch Special Study, September 2002

FITCH (2007a): Bank Failures Study 1990−2003, Fitch Special Study

FITCH (2007b): Bank Systemic Risk Report, Fitch Special Study, September 2007

GERŠL, A., HEŘMÁNEK, J. (2007): Financial Stability Indicators: Advantages and Disadvantages of their Use in theAssessment of Financial System Stability, Financial Stability Report 2006, Czech National Bank, pp. 69−79

LIŠKA, V., VINŠ, P. (2005): Rating, ISBN 80-7179-807-X, C. H. Beck, Prague

MOODY'S (2005): Defaults, Losses and Rating Transitions on Bonds Issued by Financial Institutions: 1983−2004,Moody's Special Comment, December 2005

POTTIER, S. W. , SOMMER, D. W. (1999): Property-Liability Insurer Financial Strength Ratings: Differences AcrossRating Agencies, Journal of Risk and Insurance, Vol. 66, No. 4, pp. 621−642

SŮVOVÁ, H., KOZELKOVÁ, E., ZEMAN, D., BAUEROVÁ, J. (2005): Eligibility of External Credit AssessmentInstitutions, Research and Policy Note No. 3, Czech National Bank

STANDARD&POOR'S (2007): Insurer Financial Strength Rating Definitions

Databases of rating agencies: www.moodys.com, www.fitchratings.com, www.standardandpoors.com.

76 SCORING AS AN INDICATOR OF FINANCIAL STABILITY

SCORING AS AN INDICATOR OF FINANCIAL STABILITY

Petr Jakubík, CNB, and Petr Teplý, Institute of Economic Studies, Faculty of Social Sciences, Charles University inPrague

This article presents a financial scoring model estimated on Czech corporate accounting data. Seven financialindicators capable of explaining business failure at a 1-year prediction horizon are identified. Using the modelestimated in this way, an aggregate indicator of the creditworthiness of the Czech corporate sector is thenconstructed and its evolution over time is shown. This indicator aids the estimation of the risks of this sector goingforward and broadens the existing analytical set-up used by the Czech National Bank for its financial stabilityanalyses. The results suggest that the creditworthiness of the Czech corporate sector steadily improved between2004 and 2006.

1. INTRODUCTION

Credit scoring methods are a standard part of financial institutions' risk management processes. They allowlenders to rate the creditworthiness of their potential debtors by estimating the probability of default,118 with theaim of maintaining a high-quality loan portfolio. The most common type of credit scoring used in banks for thelegal entities segment is financial scoring. In this case, companies are rated using financial indicators based ontheir accounting statements. The financial scoring process generates a score expressing the company'screditworthiness. This type of model can be applied analogously to aggregate economic data to construct afinancial stability indicator based on the creditworthiness of the non-financial sector. From the credit riskassessment perspective, the indicator can be used to complement the sectoral macroeconomic models that havebeen estimated for the Czech economy and incorporated into the banking sector stress tests (Jakubík, 2007a).

This article begins by looking briefly at the definition and estimation of scoring models (section 2). Section 3discusses the corporate financial indicators that can be used as explanatory variables for business failure. Section 4contains a description of the data used to estimate the model. The resulting estimated model is presented insection 5, and section 6 then applies the model to data for the entire sector to estimate a creditworthinessindicator for the non-financial corporations sector. The final section summarises the results.

2. SCORING FUNCTION

Scoring models play a role in the decision whether or not to provide a loan.119 In practice, this is done by comparinginformation available on the client (obtained, for instance, from the client's loan application form or track record)against information on clients to whom loans have been granted in the past and whose quality is known. Apredictive scoring model is estimated from the historical information on clients. By applying the model to knowninformation on a potential obligor, one obtains the probability that the obligor will default. The decision is madeby comparing the estimated probability of default against some threshold. A survey of these methods in thecontext of credit scoring can be found, for example, in Hand and Henley (1997) and Rosenberg and Gleit (1994).

A whole range of statistical methods can be used to construct scoring functions, among them linear regression,decision trees, neural networks and expert systems. In practice, however, logistic regression is the most commonlyused method. In this case, it is assumed that the explanatory variables multiplied by the relevant coefficients arelinearly related to the natural logarithm of the default rate (referred to as the logit − Mays, 2001).

118 Default is generally defined as the failure of an obligor to meet its obligations arising under a loan agreement. The Basel Committee on Banking Supervision (2006) defines default as a situation where at least one of the following events has taken place. The first is the situationwhere the bank finds that the obligor is unlikely to pay its credit obligations in full, without recourse by the bank to actions such as realisingsecurity. The second is the situation where the obligor is past due more than 90 days on any of its obligations. In this article, default will meanthe failure of the firm.

119 Altman (1968) and Beaver (1966) studied the prediction of corporate bankruptcy as early as the 1960s. These methods were refined in the 1980s and, in particular, the 1990s by, for example, Dimitras, Slowinski, Susmaga and Zopounidis (1999). In the Czech literature, creditscoring has been studied by, for example, Jakubík (2003).

SCORING AS AN INDICATOR OF FINANCIAL STABILITY 77

(1)

wheres - denotes the probability of default of the firm at the one-year forecast horizon,xi - denotes the financial indicators of the firm,bi - denotes the coefficients of the relevant scoring function indicators.

This equation can then be used to derive the relationship for the probability of default. This relationship can beexpressed using a logit curve.

(2)

In the case of financial scoring, financial indicators based on accounting data are considered as the explanatoryvariables. The coefficients of the function can be estimated using the maximum likelihood method (Baltagi, 2002).Owing to the large number of indicators that can be included in the model, stepwise regression is used to selectthe variables. This method involves testing various combinations of variables maximising the quality of the model.The model works with a binary dependent variable (0/1) and can be constructed for computation of either theprobability of default or the probability of non-default, depending on the definition of the independent variablein the regression. If we denote a "bad firm"120 with the value 1, the resulting score obtained from the modelcorresponds to the probability that the firm will default.121

If we assume that a large number of firms are used in order to estimate model (1), then according to the law oflarge numbers the variable s in equations (1) and (2) corresponds to the proportion of firms that default at theone-year forecast horizon. Assuming that model (2) is estimated on the set of firms to which the function will laterbe applied, the outcome of the model truly represents the probability of default. As the ratio of good to bad firmsin the sample does not usually match the real situation, and given also that accounting data from variousmoments in time are taken into consideration, the outcome of the model cannot be interpreted as the probabilityof default. In this context, variable s is usually referred to as the score expressing the riskiness or creditworthinessof the firm.122

3. FINANCIAL INDICATORS

The financial indicators used as the explanatory variables in model (2) can be broken down according to severalperspectives − for example the perspectives of lenders, shareholders or state authorities. It is important toemphasise that there is no clear consensus either in theory or in practice on the ideal method for analysing thefinancial indicators. In the Czech literature, various authors present various breakdowns of relative indicators −see, for example, Blaha and Jindřichovská (2006) and Kislingerová (2007). There is a similar lack of unity in theforeign literature − see, for example, Damodaran (2002) and McKinsey et al. (2005).

Given the primary aim of our research, namely to construct a financial stability indicator based on the predictionof business failure, we chose 22 indicators and divided them into four main groups: liquidity indicators, solvencyindicators, profitability indicators and activity indicators. The individual financial indicators are given in Table 1. Foreach indicator we also indicate its theoretical influence on business failure (positive or negative).

120 A bad firm is defined here as a firm that defaults during the period under review but was a good firm prior to defaulting. A good firm meansa firm that does not default during the period under review.

121 Some studies, conversely, denote "good firms" with the number 1. In this case, the resulting score represents the probability that the firm willnot default.

122 The figure obtained can be converted to the probability of default with the aid of a suitable transformation. Either parametric or non-parametric estimates can be used for this purpose.


The liquidity indicators explore the firm's ability to meet its short-term liabilities (r1, r2, r15 and r19) or to coverits long-term liabilities with long-term assets (r10). Generally, higher liquidity implies a lower probability of default.Persisting problems with low liquidity usually indicate problems ahead with meeting long-term liabilities (i.e. declining solvency123), which in the extreme case can result in company failure.

The solvency indicators describe the firm's ability to meet its long-term liabilities. Generally, a higher debt ratio (r3, r4 and r14) and a longer debt repayment period (r9) result in a higher probability of default. By contrast, anability of the company to generate sufficient funds for debt repayment (r5, r6, r13 and r16) and a higherproportion of internal funds (r17) reduce this probability.

The profitability indicators explain how the company generates profit and the quantity of inputs it uses to do so.Generally, higher profitability implies a lower probability of default (r7, r8, r20 and r21).

The activity indicators measure the efficiency of use of various inputs by the company. From the financial point ofview, it would be ideal if the company generated sales/profit by using the minimum amount of resources.Generally, the lower the company's efficiency, the higher its probability of default (r11, r12 and r22). The salesturnover ratio (r18) is constructed so that the value of the indicator rises − and the probability of default falls −as the volume of sales rises.

The potential influence of the individual indicators on corporate bankruptcy can be demonstrated on thefollowing simplified example.124 One classic symptom of declining solvency is when a company fails to makeefficient use of inputs (its activity indicators deteriorate). Cash flows into the firm consequently shrink, leading toa decline in the firm's ability to meet its short-term liabilities (its liquidity indicators deteriorate). Over time, thecompany proves to be incapable of generating a profit (its profitability indicators deteriorate) to cover its short-term and long-term liabilities (its solvency indicators deteriorate). The firm's liabilities exceed its assets and it goesbankrupt.

To estimate model (1), the financial indicators obtained using the relationships given in Table 1 were furthertransformed into their relative order vis-a-vis the data sample used. In this way, each indicator value wastransformed into a number lying in the range 0−1. This simple transformation makes the model estimate morerobust to outlying values of the indicators considered.

123 Liquidity is sometimes referred to as the short-term solvency of a company. 124 In this simplified example we ignore alternative ways of restoring the firm to health (e.g. corporate restructuring, debt capitalisation and so on).

Table 1 − Financial indicators

Definition Notation Expected impactRatio

Liquidity ratios

Current ratio

Quick ratio

Cash ratio

Working capital

Capitalisation ratio

current assetscurrent liabilities

cash + STi receivablescurrent liabilitiesworking capital

assetsfinancial assets

current liabilitiesfixed assets

long-term liabilities

r1

r2

r19

r15

r10

-

-

-

-

-

Solvency ratios

Leverage I

Leverage II

Leverage III

debtequity

LTii debt + LTii bondsequitydebtassets

r3

r4

r14

+

+

+


4. DATA USED

For our research we used the large database of the Czech Capital Information Agency (Česká kapitálová infor-mační agentura, ČEKIA), which contains the accounting statements (balance sheets and profit-and-loss accounts)of selected Czech firms for the period 1993−2005. Of the total of 31,612 firms in the database, 932 wentbankrupt. Since some of the accounting statements had been completed very sparsely, we focused on the recordsof firms whose main economic activity (NACE) was filled in, because for these firms most of the accounting itemswere filled in as well. In order to estimate the scoring function, from the firms that went bankrupt we initiallyselected only those for which there was accounting data one year prior to the declaration of bankruptcy. Therewere 151 such firms.125 Then, for the sample of firms that did not fail in the period under review we selected onlythose for which we had accounting statements for at least two consecutive years.126 The data sample for the

125 We excluded from our analyses those firms which underwent composition. There were only nine such cases in the database. Unlike bankruptcy, composition is not associated with the dissolution of the legal entity (Jakubík, 2007b).

126 To estimate the scoring function we need to have corporate accounting data for two consecutive years. The first period is used for estimatingthe function and the second for identifying the quality of the firm (failed, healthy). If no accounting data are available for the following period, we are unable to determine the quality of the company in question.

Table 1 − Financial indicators −− continued

i Short-term ii Long-term

Definition Notation Expected impactRatio

Solvency ratios

Debt payback period

Interest coverage

Cash-flow I

Cash-flow II

No credit interval

Retained earnings

LTii debt + STi debtoperating profit+interest expenses+depreciation

operating profit + interest expensesinterest expenses

net profit + depreciation(debt-reserves)/365

net profit + depreciationdebt/365

money + STi payables + LTii payables)operating expensesretained earnings

assets

r9

r5

r6

r13

r16

r17

+

-

-

-

-

-

Profitability ratios

Gross profit margin

Return on assets

Return on equity

Net profit margin

operating profitsales

operating profitassets

net profitequity

net profitsales

r7

r8

r20

r21

-

-

-

-

Activity ratios

Average receivable collection period

Inventory ratio

Sales turnover

Payables ratio

receivablessales/365

inventoriessales/365

salesassets

STi payablessales/365

r11

r12

r18

r22

+

+

-

+


estimation of the model was constructed so as to best capture the true data structure. Usually, however, a largerproportion of bad firms than exists in reality is included in the sample so that the good and bad firms can bedistinguished using statistical methods. Sometimes a sample containing the same number of good and bad firmsis used (Wezel, 2005). Generally, the good firms are chosen so as to be as similar as possible to the bad onesaccording to selected criteria, for example size as measured by assets, number of employees or sales.127 We alsorandomly selected accounting periods for which statements were available for the immediately succeedingaccounting period. In this way we made sure that the firm in question did not fail in the year following the periodunder review. In all, 606 good firms were ultimately selected using this procedure. The data sample thus containeda total of 757 firms, which were divided into two categories according to whether they went bankrupt in theperiod following the period for which the accounting data were selected for the company in question. Accordingto the econometric literature, when the event of interest is rare, logistic regression underestimates the influenceof the characteristics on the event, so an artificial sample is generated and the estimated values are furthertransformed so that they match the incidence in the population.128

Table 2 shows the breakdown of the data in the database on the selected data sample by accounting period andfirm quality (good/bad). In the total data sample, moreover, there exists a set of firms for which we are unable todetermine the quality in the given year (indeterminate firms). These are firms for which accounting statements forthe following year are not available. Although the database contained accounting data for the period 1993−2005,in the final year it is no longer possible to determine the firm's quality. For this reason, the selected data sampledoes not cover 2005.

Table 3 shows the breakdown of firms by size in the data sample. This is based on corporate assets and conformsto the European Commission categorisation.129 Nonetheless, we should mention that the European Commissionalso offers other enterprise size categorisations (according, for example, to number of employees or sales).130 The

127 A summary of the methods can be found, for example, in Heckman et al. (1997). 128 For the estimation of the scoring function, an alternative sample constructed in the same way but with a new random selection for the good

firms was used in the robustness tests − see section 5, where we discuss the results of the model. 129 Commission Regulation (EC) No 70/2001 as amended by No 364/2004. The enterprise size boundaries were converted from EUR to CZK using

the approximate exchange rate 1 EUR = 30 CZK. 130 The Czech Statistical Office also uses a breakdown by number of employees.

Table 2 − Breakdown of data sample by accounting period and firm quality*

199319941995199619971998199920002001200220032004Total**

Total data

Total Undefined firms Bad firms Good firms Total Bad firms Good firms

Used data sample

980 89 1 890 1 1 01.824 53 0 1.771 4 0 45.606 147 0 5.459 13 0 137.023 1.032 9 5.982 53 9 447.056 1.261 15 5.780 50 15 356.802 1.028 12 5.762 48 11 377.541 1.307 25 6.209 69 25 447.377 3.094 18 4.265 62 17 455.660 1.536 5 4.119 40 5 357.869 956 8 6.905 57 8 49

22.264 4.420 25 17.819 110 25 8518.989 18.490 35 464 250 35 21598.991 33.413 153 65.425 757 151 606

Source: ČEKIA and authors' calculations* A bad firm means a firm that went bankrupt at the one-year horizon, whereas a good firm for the given period means a firm that did not go bankrupt the

following year.** The "Total" row contains the number of observations for the given set of firms. On the full data sample this figure does not equal the total number of firms,

because in the selection each company is monitored for several accounting periods.


enterprise size definition chosen by us and used in Table 3 was based on the available data, which were part ofthe data source used. The source contained corporate assets, and not numbers of employees. Sales informationdid form part of the database, but had been filled in for only some companies, so it could not be used. Under thedefinition we used, micro-enterprises with assets not exceeding CZK 60 million have the largest representation inthe data sample, while large enterprises with assets exceeding CZK 1,290 million have the lowest representation.However, large enterprises account for more than 80% of the aggregate assets of the firms represented in thesample.

5. RESULTS OF THE MODEL

The resulting model (3) confirmed the relationships between the liquidity, solvency, profitability and activityindicators and business failure. The best statistical properties were shown by the model containing sevenstatistically significant indicators (of the 22 considered in all). These included three solvency indicators (leverage Iand II and interest coverage), two profitability indicators (return on equity and gross profit margin), one liquidityindicator (cash ratio) and one activity indicator (inventory ratio). The resulting model takes the following form:

(3)

where

score - expresses the risk of the firm, which is linked to the probability that the firm will go bankrupt at the one-year horizon,

ri - denotes the individual financial indicators of the firm,bi - denotes the coefficients of the relevant scoring function indicators,* - denotes the relative order operator in per cent, which returns the relative order of the value of a given

indicator for a given firm vis-a-vis the full data sample used to estimate the model.131

131 The relative order operator returns a number in the range of 0−1. It is analogous to seeking a quantile on the given data sample, except thatthe value for which we are seeking the position in the given sample is not part of the sample. In practice, we calculate the value of a givenfinancial indicator, such as the cash ratio, and seek the two closest indicator values in the data sample between which the value sought lies.From the relative order of these two values we calculate the relative order for the sought value by linear interpolation. If, for example, the cashratio takes the value 0.2, the relative order operator for it is calculated by linear interpolation of the relative order of the two closest values to0.2 occurring in the data sample used for the estimation of the model, namely 0.1996 and 0.2015, whose relative orders are 0.5733 and0.5746. We then obtain the resulting relative order value using the following relationship:

This means that in the original data sample on which the model was estimated, 57.36% of the values of this indicator are less than 0.2.

Table 3 − Description of data sample used

Micro firmsSmall firmsMedium firmsLarge firmsTotal

Good firms

TypeAssets

(CZK millions)Number of

firms

Share according to number

of firms (%)

Share according to assets

of firms (%)

Number of firms

Share according to number

of firms (%)

Share according to assets

of firms (%)

Bad firms

< 60 292 48.2% 0.8% 70 46.4% 1.0%61-300 138 22.8% 3.5% 36 23.8% 5.4%

301-1.290 90 14.9% 10.8% 24 15.9% 14.7%>1.291 86 14.2% 84.9% 21 13.9% 78.9%

- 606 100.0% 100.0% 151 100.0% 100.0%

Source: ČEKIA and authors' calculations

0.5733. + 0.5746. = 0.5736, i.e. 0.2* = 0.5736.0.2015-0.2

0.2015-0.1996

0.2-0.1996

0.2015-0.1996


As the model is based on the relative order of the indicators in the sample, the estimated coefficients of thefunction express their relative importance. The larger is the indicator's coefficient (in absolute terms), the larger isits weight in the scoring function.132 From this perspective, interest coverage, cash ratio and leverage I appear tobe the most important indicators (Table 4).

The estimated scoring model confirmed our expectations regarding the impact of the individual indicators onbusiness failure. It is clear that a higher debt ratio increases the probability of default (see leverage I and II),whereas a higher ability to repay debts (see the interest coverage) reduces this probability. Likewise, higherprofitability (see gross profit margin and return on equity) and higher liquidity (see cash ratio) increase the financialstability of the firm and reduce its probability of default. By contrast, lower efficiency (see inventory ratio) implieslower financial stability of the firm.

Although the model confirmed some of the expected results, for example that solvency and liquidity ratios are themost important for predicting corporate bankruptcy, one surprising result is the importance of inventories, ascontained in the inventory ratio (i.e. the number of days a company has goods in stock in the form of inventories).The higher this indicator is, the longer goods lie in the company's store and the less saleable its inventories are.133

One possible explanation for the importance of this indicator is the high stock of unsaleable inventories typical ofbusinesses heading towards bankruptcy. This argument is supported by the fact that the total liquidity indicator,which includes inventories in current assets, proved to be insignificant. Conversely, the cash ratio, which does notinclude inventories in current assets at all, appears to be significant. This implies that the saleability of inventories− among other indicators − plays an important role in the prediction of corporate bankruptcy.

The aim of the scoring model is to correctly separate good and bad firms. This property expresses the quality ofthe estimated function. To measure it, one can use the Gini coefficient, for example. The value of this coefficientshould be as close as possible to 1, which would mean a 100% ability to separate firms in terms of their qualityusing the scoring function. The quality of the model can be demonstrated graphically by means of a histogram(Chart 1) or a Lorenz curve (Chart 2).

132 The relative order operator applied to the individual financial indicators used in the scoring function ensures that the model is robust to extremevalues.

133 Nevertheless, we should point out that different industries display different inventory ratios. For example, this indicator is high for ship manufacture, but very low for retail trade.

Table 4 − Estimated scoring model

ConstantLeverage ILeverage IICash-flow IGross profit marginInventory ratioCash ratioReturn on equity

-SolvencySolvencySolvencyProfitabilityActitvityLiquidityProfitability

-r3r4r5r7

r12r19r20

b0

b1

b2

b3

b4

b5

b6

b7

2.41922.57791.7863

-3.4902-2.41721.7679

-3.3062-2.2491

0.92890.37880.57271.00050.48020.40330.42460.5621

0.0092070.0000000.0018130.0004860.0000000.0000120.0000000.000063

Variable TypeIndicatorNotation

CoefficientNotation

CoefficientStandard

ErrorSignificance

Source: Authors' calculations


Chart 1 shows the firm distribution of the data sample used according to score and according to whetherbankruptcy occurred. The blue columns express the percentage of good firms and the red columns the percentageof bad firms for each score interval. The ideal situation would be if all the bankrupt firms were assigned a scoreof 1 and all the healthy ones a score of 0. This, however, does not happen in practice, as we are unable to observethe complete characteristics of the firm and so we are working with imperfect information. This implies that thefunction cannot fully separate the firms according to their quality. There is always a set of bad firms that areclassified as good ones, and vice versa. The aim is to keep such cases to a minimum.

Chart 2 depicts the cumulative distribution of the scores of good and bad firms. In the ideal case, guaranteeing a100% rate of separation, this curve would take the form of a right angle. From the Lorenz curve one can computethe "Gini coefficient" as the ratio of the area enclosed by the green curve and the black diagonal and the totalarea below that diagonal. The generally accepted Gini coefficient for this type of model fluctuates above 60%depending on the data used and the purpose of the scoring (Mays, 2001). With a Gini coefficient of 80.41%, ourestimated model satisfies the requirement of a sufficient rate of separation of the firms on the data sample used.

The estimation of the model for the alternative data sample, constructed according to the same rules as the sampleused, confirmed that our estimate is sufficiently robust. The robustness of the model was also tested on anotheralternative data sample consisting of good clients selected entirely at random, and their representation accordingto the breakdown by assets was different from both the alternative and original data samples. In this case, a slightchange was made to the model (two of the seven indicators were replaced with others134), but when the model wasapplied to aggregate data on financial corporations (discussed in section 6), similar results were obtained (theresulting score was different owing to a different ratio of good to bad clients in the sample, but the time profile ofthe score was similar). The quality of the model as measured by the Gini coefficient was also almost identical.

6. USE OF THE MODEL TO ASSESS THE FINANCIAL STABILITY OF THE ECONOMY

Financial scoring is routinely used to assess the creditworthiness of individual firms. If we have aggregated datafor the whole non-financial sector, we can imagine this sector as one large hypothetical firm with an aggregatedbalance sheet. Alternatively, given the use of relative indicators only, we can view the aggregated indicators ascharacteristics of the average firm in the sector. Assuming a degree of homogeneity, the estimated model can beapplied to the aggregated indicators of non-financial corporations. If the situation in the sector takes a turn for

134 Retained profit and cash flow I were replaced with gross profit margin and interest coverage.

0

5

10

15

20

25

30

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Good Bad

00.10.20.30.40.50.60.70.80.9

1

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Bad

Goo

d

Chart 1 − Histogram of estimated scoring function Chart 2 − Lorenz curve of estimated scoring function

Source: Authors' calculations Source: Authors' calculations


the worse, the financial indicators of firms will deteriorate on average. This will be reflected in a falling score ofthe average representative firm. However, the scope and inhomogeneity of the sample of firms on which themodel was estimated place some limitations on the model. We could get better results by decomposing thesample into several more homogeneous segments and then estimating the model for these groups of firmsseparately. In the ideal scenario, we would decompose the firms by size and area of economic activity. Owing tothe small number of bad firms in the data source used, however, this is not possible.

An aggregated balance sheet can be obtained for Czech firms from the publicly available data of the CzechStatistical Office, which has data containing the economic results of non-financial corporations. This data ispublished in a sufficiently detailed structure (to enable the construction of the seven aforementioned indicatorsincluded in the model) only for corporations with 100 employees or more. The seven indicators obtained in thisway (r3, r4, r5, r7, r12, r19 and r20) are substituted into equation (3) to give an aggregated score representing thelevel of risk of the entire sector.

The resulting score was computed for 2004−2006.135 The value of the creditworthiness indicator (the 1-score) for2004−2006 (see Chart 3) can be interpreted as the creditworthiness of the non-financial sector for the one-yearforecast horizon. This indicator is related directly to the probability of default of the corporate sector. By contrastwith the original data sample, the model is only applied to data on firms with 100 employees or more, but one canget some idea of the evolution of the corporate sector over time. Given the aforementioned limitation, the resultingscore is probably underestimated and thus the creditworthiness is overestimated, owing to the higher level of riskof the small enterprises excluded from the aggregate data. For financial stability purposes, however, the dynamicsof this indicator over time are more important than its absolute level. The results suggest a steady improvement inthe creditworthiness of the non-financial sector between 2004 and 2006 in line with the positive macroeconomictrend. Although there was a slight decline in return on equity and the gross profit margin in 2005, the positivedevelopment of the other five indicators (a falling debt ratio, rising liquidity, increasing interest coverage and adecreasing inventory ratio) outweighed this effect and the resulting creditworthiness score increased slightlycompared to 2004. A positive shift and a reduction in the risk of the sector occurred in particular in 2006, whichsaw improvements in five out of the seven financial indicators studied (the only deteriorations were recorded byleverage I and II). The biggest improvements were shown by interest coverage (a year-on-year improvement of24.2% to 11.78), cash ratio (a year-on-year increase of 24.2% to 0.385) and return on equity (a year-on-yearimprovement of 10.7% to 0.105). The results of the model are consistent with the conclusions contained in the2006 Financial Stability Report, according to which 2006 was an extraordinarily successful year for the largecorporations sector and the outlook for 2007 was favourable (CNB, 2007). The constructed indicator offers a morecomprehensive aggregate view of the riskiness of the sector as a whole going forward.

135 Since some of the balance sheet items needed to calculate the necessary indicators were not monitored until after 2004, these values for 2004were estimated from the available data for 2004 and 2005.

0.97110.9716

0.9734

9.99.5

10.5

0.968

0.969

0.970

0.971

0.972

0.973

2004 2005 20069.00

9.50

10.00

10.50

11.00

11.50

12.00

12.50

Creditworthiness (left-hand scale, 1-score) Return on equity (right-hand scale, %)

Chart 3 − Creditworthiness of the non-financial corporations sector

Source: Authors' calculations and CZSO

7. CONCLUSIONS

Financial scoring is a method used to assess the creditworthiness of obligors and thus is frequently used by lenderswhen deciding whether or not to provide credit products. This study showed that it is possible to use thesetraditional methods to monitor the financial stability of the corporate sector. Using accounting data on Czechfirms, a scoring model based on seven financial indicators was estimated using logistic regression. By applying thismodel to the aggregate financial results of non-financial corporations, the scores of the Czech corporate sectoras a whole − corresponding to its level of risk for the one-year forecast horizon − were calculated for 2004−2006.The results of our study suggest that the creditworthiness of the Czech non-financial corporate sector improvedbetween 2004 and 2006. This indicator will be incorporated into the quantitative system used by the CzechNational Bank to assess financial stability. The calculated score will be used each year as auxiliary information forevaluating the probability of the corporate sector running into difficulties at the one-year forecast horizon.

REFERENCES

ALTMAN, E. I. (1968): Financial Ratios, Discriminant Analysis and the Prediction of Corporate Bankruptcy, Journalof Finance, Vol. 23, No. 4, pp. 589−609

BALTAGI, B. D. (2002): Econometrics, Springer

BASEL COMMITTEE ON BANKING SUPERVISION (2006): International Convergence of Capital Measurement andCapital Standards: A Revised Comprehensive Framework, Bank for International Settlements

BEAVER, W. (1966): Financial Ratios as Predictors of Failure, Journal of Accounting Research, Vol. 4, pp. 71−102

BLAHA, Z. S., JINDŘICHOVSKÁ, I. (2006): Jak posoudit finanční zdraví firmy: finanční analýza pro investory:bankéře, brokery, manažery, podnikatele i drobné akcionáře, 3rd edition, Management Press


DAMODARAN, A. (2002): Investment Valuation: Tools and Techniques for Determining the Value of Any Asset,2nd Edition, John Wiley & Sons

DIMITRAS, A. I., SLOWINSKI, R., SUSMAGA, R., ZOPOUNIDIS, C. (1999): Business Failure Prediction Using RoughSets, European Journal of Operational Research, Vol. 114, No. 2, pp. 263−280

HAND, D., HENLEY, W. (1997): Statistical Classification Methods in Consumer Credit Scoring: A Review. Journal ofthe Royal Statistical Society, Vol. 160, No. 3, pp. 523−541

HECKMAN, J., ICHIMURA, H., TODD, P. (1997): Matching as an Econometric Evaluation Estimator: Evidence fromEvaluating a Job Training Program, Review of Economic Studies, Vol. 64, No. 4, pp. 605−654

JAKUBÍK, P. (2003): Úloha skóringu při řízení kreditního rizika, Acta Oeconomica Pragensia − Finanční krize, VŠEPraha, vol. 1, pp. 147−157

JAKUBÍK, P. (2007A): Macroeconomic Environment and Credit Risk, Czech Journal of Economics and Finance, Vol. 57, Nos. 1 and 2, pp. 60−78

JAKUBÍK, P. (2007B): Exekuce, bankroty a jejich makroekonomické determinanty, IES Working Paper 2007/29

KISLINGEROVÁ, E. (2007): Manažerské finance, 2nd Edition, C. H. Beck

MAYS, E. (2001): Handbook of Credit Scoring, Glenlake Publishing Company, Ltd.

MCKINSEY ET AL. (2005): Valuation − Measuring and Managing the Value of Companies, 4th Edition, John Wiley & Sons

ROSENBERG, E., GLEIT, A. (1994): Quantitative Methods in Credit Management: A Survey, Operations Research,Vol. 42, No. 4, pp. 589−613

WEZEL, T. (2005): Determinants of Foreign Direct Investment in Emerging Markets: An Empirical Study of FDIFlows from Germany and its Banking Sector, Peter Lang GmbH, Frankfurt am Main


86 COMPETITION AND EFFICIENCY IN THE CZECH BANKING SECTOR

COMPETITION AND EFFICIENCY IN THE CZECH BANKING SECTOR

Anca Podpiera, CNB*

This article offers empirical evidence on the evolution of competition in the banking sector and its relationship tothe cost efficiency of banks in the Czech Republic between 1994 and 2005. First, we measured the level andevolution of competition and cost efficiency. Competition was measured with the Lerner index on the loan market,using quarterly data on loan prices. Then we investigated the relationship and causality between competition andefficiency using the Granger-causality test. This supported a negative causality running from competition toefficiency (the "banking specificities" hypothesis). By contrast, our results reject the "quiet life" hypothesis for theCzech banking sector, according to which competition should positively influence efficiency.

1. INTRODUCTION

Banking competition is expected to provide welfare gains by reducing monopoly rents and cost inefficiencies. A higher degree of banking competition should result in a lower monopoly power of banks, and therefore a decrease in banking prices. Since investment is particularly sensitive to a decrease in commercial interest rates,a reduction of monopoly rents should consequently impact positively on investment and economic growth. Theseexpected gains are a particularly major issue for countries in which bank credit represents the largest source ofexternal finance for companies, as is the case in the Czech Republic (Reininger et al., 2002). Heightenedcompetition should also encourage banks to reduce their operating costs, i.e. their cost inefficiencies. This latterchannel is particularly promising in terms of welfare gains, as the order of magnitude of the banking sector costinefficiencies has been estimated in the past to average between 30% and 50% in European transition countries(e.g. Hasan and Marton, 2001; Fries and Taci, 2005) and 40% in the Czech Republic (Podpiera and Podpiera,2005).136 However, some studies emphasise some potential negative effects of growing banking competitionthrough excessive risk-taking by banks, which may hamper financial stability (Allen and Gale, 2004; Carletti andHartmann, 2002).

The aim of this research is twofold. First, we provide empirical evidence on the level and evolution of competitionin the Czech banking sector between 1994 and 2005. In previous studies for the Czech banking sector,competition has been measured using concentration indices such as the Herfindahl index, with higherconcentration signalling lower competition and vice versa. This approach is based on the traditional IndustrialOrganisation (IO) literature, which uses structural tests to assess banking competition based on the SCP modelderived in Bain (1951). The SCP hypothesis argues that greater concentration causes less competitive bankconduct and leads to greater profitability (meaning lower performance in terms of national welfare). Accordingto this theory, competition can be measured by concentration indices such as the market share of the five largestbanks, or by the Herfindahl index. These tools were applied until the 1990s. However, they suffer from the factthat they infer the degree of competition from indirect proxies such as market structure or market shares.

The new empirical IO approach is based on non-structural tests to circumvent the problems of measuringcompetition by the traditional IO approach. The new empirical IO theory determines banks' conduct directly.Furthermore, it allows us to consider the actual behaviour of banks by taking contestability into account. Indeed,as observed by Claessens and Laeven (2004), the actual behaviour of a bank is related not only to marketstructure, but also to barriers to entry, influencing the likelihood of the entry of new competitors and thereforethe behaviour of incumbents.

In this study we measure competition with the Lerner index, using data on output prices.137 The Lerner index hasbeen computed in several empirical studies on banking competition (e.g. Angelini and Cetorelli, 2003; Fernandez

* This article is based on Podpiera, A., Weill, L., Schobert, F. (2007): Market Power and Cost Efficiency in the Czech Banking Sector, CNB WorkingPaper 5/2007.

136 The numerical data describe the excess costs of the average bank relative the most efficient bank for producing the same bundle of outputs.137 The most commonly applied tool for assessing competition emanating from the new empirical IO approach is the Rosse-Panzar model. This non-

structural test is based upon the estimation of the H-statistic, which aggregates the elasticities of total revenues to input prices. It has beenapplied in Western European countries by several authors (Bikker and Haaf, 2002; Weill, 2004; Gelos and Roldos, 2004). Gelos and Roldos (2004)includes three transition countries (the Czech Republic, Hungary and Poland) and concludes in favour of monopolistic competition in these threecountries' banking markets and also of the absence of a significant change in banking competition between 1994 and 1999.

COMPETITION AND EFFICIENCY IN THE CZECH BANKING SECTOR 87

de Guevara et al., 2005). In this study we focus exclusively on the loan market, which represents by far thegreatest share of assets for the Czech banking sector. We are therefore able to measure the degree of monopolypower for each bank in the loan market and the evolution of competition over the period 1994−2005.

The second aim is to investigate the relationship and causality between competition and efficiency. Indeed, in spiteof the commonly accepted view favouring a positive relationship, there does exist empirical literature supportinga negative link (Berger, 1995; Goldberg and Rai, 1996; Weill, 2004). The theoretical literature provides argumentsfor both directions of the causality. The intuitive "quiet life" concept (Hicks, 1935) suggests that competitionpositively influences efficiency. In other words, this concept suggests that monopoly power allows a relaxing ofefficiency efforts.138 By contrast, the "efficient-structure" hypothesis, proposed by Demsetz (1973), predicts anegative impact of efficiency on competition, as the most efficient banks would benefit from lower costs andtherefore higher market shares.

On the other hand, the specificities of banking competition may give rise to a negative impact of competition onefficiency. The theoretical literature in banking suggests that imperfect competition in banking markets may resultfrom the information asymmetries between bank and borrower in credit activity. As a consequence, banks haveto implement some mechanisms to resolve the resulting problems such as adverse selection and moral hazard.One way out is the implementation by the bank of a customer relationship, meaning a long-term repeatedrelationship, to gain better information on the borrower and reduce the information asymmetries. According toDiamond (1984), banks have a comparative advantage in the ex post monitoring of borrowers, in comparison toinvestors, because of the existence of economies of scale resulting from their role of delegated monitor. In thiscase, competition may make it impossible to realise such economies of scale. As a consequence, competition mayincrease monitoring costs and potentially reduce the length of the customer relationship, further decreasing thecost efficiency of banks. In other words, the specificities of the banking industry provide some additionalarguments in favour of a negative relationship between competition and cost efficiency. This assumption will becalled the "banking specificities" hypothesis in the following text. This hypothesis may be more relevant intransition economies than in developed market economies. Indeed, banks suffer more from informationasymmetries in transition countries, because of uncertainties of accounting information and the relative lack ofcredit risk analysis know-how of bank employees, owing to the short history of the market economy.

We perform Granger-causality-type estimations in order to get information on the sense of the causality betweencompetition and efficiency in banking. An analysis of this type will enrich the debate on the conflicting hypothesesdescribed above. Such an analysis is also important to provide the normative implications of competition policy inthe banking industry. Specifically, a negative relationship between competition and efficiency would mean a trade-off between these two objectives.

2. METHODOLOGY

Measurement of Competition

The Lerner index is defined as the difference between the price of output (loans) and marginal cost divided byprice. The price of loans is computed as "Total interest revenues" divided by "Total net loans". "Total net loans"represents "Total loans" from which non-performing loans have been subtracted.139 The marginal cost is based on the estimation of the cost function. We estimate a cost function with one output and three input prices. Onecost function is estimated for each year by introducing fixed effects for individual banks. We impose the restrictionof linear homogeneity in input prices by normalising total costs and input prices by one chosen input price. Thecost function is specified as follows:

138 This argument is summarised in the famous sentence from Hicks: "The best of all monopoly profits is a quiet life".139 Revenues are not likely to come from non-performing loans, so the inclusion of non-performing loans in the Lerner index calculation would

understate the price for banks having significant proportions of non-performing loans.


(1)

where TC denotes total costs, y loans, w1 the price of labour, w2 the price of physical capital, and w3 the price ofborrowed funds. The indices for each bank have been dropped from the presentation for the sake of simplicity.

The estimated coefficients of the cost function are then used to compute the marginal cost.

The marginal cost can be expressed as follows:

(2)

The derivative of the logarithm of the total cost with respect to the logarithm of output is computed using thecost function specified in equation (1):

(3)

Measurement of Efficiency

We measure efficiency on the basis of cost efficiency, meaning that we measure how close a bank's cost is to whatthe most cost-efficient bank's cost would be for producing the same bundle of outputs. It then provides informationon losses in the production process and on the optimality of the chosen mix of inputs. Several techniques have beenproposed in the literature to measure efficiency with frontier approaches. In this study, we adopt a distribution-freeapproach (DFA), in this way circumventing the main critique attached to the widely used stochastic frontierapproach, namely its reliance on distributional assumptions. The DFA assumes the cost function TC = f(Y,P) + εwhere TC represents total cost, Y denotes output, P is the vector of input prices and ε is the error term. Accordingto this cost function the efficiency of each bank is constant over time and the random error for each bank tendsto cancel out over time. Bauer et al. (1998) distinguish three different techniques through which DFA could beimplemented in practice. In this study, we focus on DFA-P WITHIN, which is a fixed-effects model that estimatesinefficiency from the value of a bank-specific dummy variable; each bank's efficiency is then computed as thedeviation from the most efficient bank's intercept term. We estimate the translog cost function presented inequation (1) for each year (four quarters), where we assume that the random error cancels out over the four quarters.

Testing the Relationship between Competition and Efficiency

We analyse the link between competition and efficiency in the Czech banking industry in a Granger-causalityframework, formally specified in equations (4) and (5) as follows:

(4)

(5)

where y represents "Efficiency" and x the "Lerner index". fi represents the bank's "individual effect".


Efficiency and Lerner index are the yearly averages of the cost efficiency score and the Lerner index respectively. iand t represent the indices for the bank and the time (year) respectively. Each dependent variable is regressed onits lags and on those of the other variable. We use yearly averages in order to be able to capture the genuineeffect, if any, of competition on efficiency and vice versa. In other words, we believe that it takes time for theeffect of competition on efficiency and vice versa to become apparent, hence such an effect could be revealed byanalysing yearly data rather than quarterly data, which are obviously more volatile. Following Berger and De Young(1997) we adopt a maximum lag of four years.

Having at our disposal a panel, we do not employ a standard Granger-causality analysis but resort to panel-specific methodology to estimate the dynamic equations (4) and (5). To estimate the dynamic equationsrepresented in (4) and (5) we employ the Generalised Method of Moments (GMM) as designed by Arellano and Bond (1991).140

3. DATA AND VARIABLES

In the analysis we used monthly data reported to the Czech National Bank (CNB) for all Czech commercial banks141

during the period 1994−2005, and transformed them into quarterly data.142 Two approaches are proposed in thebanking literature for the definition of inputs and outputs. The intermediation approach assumes that the bankcollects deposits to transform them, using labour and capital, into loans. The production approach views the bankas a production unit using labour and capital to produce deposits and loans. As our focus is on lending activity,we adopted the intermediation approach.

One output − loans − is adopted in the cost function and the cost efficiency frontier. The inputs include labour,physical capital and borrowed funds. The price of labour is measured by the ratio of personnel expenses to thenumber of employees. The price of physical capital is defined as the ratio of expenses for physical capital to totalfixed assets. The price of borrowed funds is measured by the ratio of expenses for borrowed funds to borrowedfunds. Total costs are the sum of expenses for personnel, physical capital and borrowed funds. The price of loansis computed using the ratio of interest received on loans to net loans. Summary statistics for the period1994−2005 are reported in Table 1.

140 Attanasio et al. (2000) mention that most studies seeking Granger-causality type estimation with fixed effects use estimators such as thoseproposed by Holtz-Eakin, Newey and Rosen (1988) and Arrelano and Bond (1991) (hereinafter "AB").

141 Mortgage banks are not included, as their production function most likely differs from that of commercial banks.142 We performed a careful investigation of the data to find and drop outliers. For the failed banks, the observations for the year of failure were

dropped, as the data for the quarters preceding the failures were mostly chaotic. Furthermore, for each bank and for each year, we tried tohave complete data for all four quarters. The result is an unbalanced panel.

Table 1 − Descriptive statistics

Median Mean Standard Deviation

OutputLoans (CZK billions)Input pricesPrice of labour (CZK thousands)Price of physical capitalPrice of borrowed fundsOther characteristicsAssets (CZK billions)Total costs (CZK millions)Price of loans

14.4 53.9 92.8

85.9 116.3 93.70.09 0.137 0.122

0.012 0.015 0.011

20.12 81.09 146.3305.4 981.8 1 727.80.021 0.023 0.0122

N=1110 observations.


4. RESULTS

The Evolution of Banking Competition

The results of the computation of the Lerner index for each year are displayed in Table 2. The Lerner index is aninverse measure of competition, i.e., a greater Lerner index means lower competition. Table 2 also shows theHerfindahl indices143 to allow comparison of our measure of competition with one used routinely in practice.

The fact that the results show a negative figure for the Lerner index in 1997 make interpretation of the trenddifficult for the period 1995−1998. The drastic increase in the Lerner index from 1998 to 1999 was to some extenttriggered by a decrease in banks' marginal costs related to a decline in interest rates on the interbank market after1998. The obvious increase in competition during the period 1999−2002 can be attributed to the entry of foreignbanks into the Czech banking industry, which considerably increased from 1999 onwards with the launch of theprivatisation of major banks. The subsequent (2003−2005) decrease in our measure of competition contradictsthe common belief of rising competition in banking. This might actually have been a result of a temporary absenceof the acute threat of a new competitor entering the Czech banking market. As all the big banks were nowprivatised, and as there was already a relatively high number of branches of foreign banks active in the market,the threat of entry of a new bank seemed very low. Consequently, the competitive pressures on banks werelimited. At the same time, we have to recall that our measure of competition does not account for the riskinessof banks' products. The price of output is the average for all types of loans regardless of their riskiness. The risein the Lerner index during the period 2002−2005 may have been due to some extent to the fact that after 2002banks offered a wider spectrum of products, some of them relatively riskier and pricier.

According to the Herfindahl index, concentration fell continuously from 1994 to 2000 and then strongly increasedfrom 2000 until 2002, followed by a slight decrease between 2003 and 2005. Our measure of competition andthe Herfindahl index display common turning points: minimum competition in 2000 and maximum competitionin 2002. Also, the period 2003−2005 is characterised by decreasing competition on both scales, although thedecline is rather smaller in the case of the Herfindahl index.

143 The Herfindahl index is the sum of the squares of the market shares of the individual market participants. It ranges from 0 to 10 000, with ahigher index signalling higher concentration.

Table 2 − Lerner indices and Herfindahl indices

199419951996199719981999200020012002200320042005

Lerner indices

No. of observations Median Mean Standard DeviationHerfindahl

indices

87 60.13 59.01 30.97 1381.78110 16.94 13.6 49.48 1233.4799 14.73 2.46 71.12 1204.91

106 -14.38 -26.88 83.67 1150.3386 8.77 10.94 24.26 1045.2699 32.16 30.76 31.73 1002.98

100 30.37 31.11 23.96 951.8992 24.4 29.12 24.79 1071.0392 17.1 17.03 27.7 1321.1888 50.95 43.44 30.93 1285.3575 55.11 45.74 27.66 1250.7076 44.8 42.09 26.67 1232.43

All indices are in per cent.Note: The negative figure for the year 1997 probably comes from the fact that on average the MC was higher than the price of loans that year. This was due tothe high interbank rates triggered by the financial turmoil in 1997.


The results of the GMM estimation of the dynamic equations represented in (4) and (5) are displayed in Table 3.The Sargan test and the first- and second-order auto-correlations in the differenced residuals (AR1 and AR2) satisfythe necessary conditions specified in Arellano and Bond (1991). The table reports the coefficients of the lags ofthe dependent and independent variables. Of primary interest for our hypothesis are the coefficients of the lag ofthe independent variable. For both equations (4) and (5), we test the hypothesis that δ1 = δ2 = .... = δm are equalto zero, which signals whether the independent variable Granger-causes the dependent variable. The sum of thesecoefficients, which gives an overall measure of the effect on the dependent variable, is also presented for anassessment of the sign of the relationship.

The results show that the Lerner index positively Granger-causes efficiency − hence competition negativelyGranger-causes efficiency − but efficiency does not Granger-cause competition.144 This result is consistent with the"banking specificities" hypothesis, according to which greater competition should reduce the cost efficiency ofbanks through increased monitoring costs.

Our work thus supports the literature regarding the trade-off between banking competition and financial stability(Allen and Gale, 2004). Our analysis provides another channel of transmission for the negative effects ofcompetition on financial stability, namely the negative effects of competition on the cost efficiency of banks.

144 In the equation explaining Efficiency the coefficients of the lags of the Lerner index are jointly statistically different from zero (Prob > chi2 =0.0000). In the equation explaining the Lerner index, the coefficients and the lags of Efficiency are not jointly different from zero (Prob > chi2= 0.3629).

Table 3 − Granger-causality tests

InterceptEfficiencyt-1

Efficiencyt-2

Efficiencyt-3

Efficiencyt-4

Efficiencyt-1= Efficiencyt-2== Efficiencyt-3= Efficiencyt-4=0∑ AR Efficiency coefficientsLernert-1

Lernert-2

Lernert-3

Lernert-4

Lernert-1= Lernert-2= Lernert-3== Lernert-4=0∑ AR Lerner coefficientsp- value AR1/AR2 p- value Sargan Number of observations

Dependent variable:Efficiencyt

Coefficient Std. err. Coefficient Std. err.

-0.06*** 0,011 0.06*** 0.02-0.6*** 0,12 0.11 0.150.05 0,12 0.28* 0.17-0.18** 0,09 -0.11 0.140.05 0,09 -0.05 0.14chi2( 4) = 32.94 chi2( 4) = 4.33Prob > chi2 = 0.0000 Prob > chi2 = 0.3629-0.69*** 0.24 0.24 0.320.2*** 0.07 -0.33*** 0.110.29*** 0.08 -0.17 0.120.29*** 0.08 -0.15 0.110.12** 0.06 -0.12 0.10chi2( 4) = 32.69 chi2( 4) = 11.99Prob > chi2 = 0.0000 Prob > chi2 = 0.01750.898*** 0.16 -0.77*** 0.240.05 / 0.13 0.000 / 0.240.003 0.041085 1085

Dependent variable:Lernert

*, **, *** denote estimates significantly different from zero at the 10%, 5% and 1% levels respectively. AR means auto-regressive lag.


5. CONCLUSIONS

This research focuses on the relationship between competition and efficiency in the Czech banking sector duringthe economic transition period. Our measure of competition shows an absence of increased competition in theCzech banking market between 1994 and 2005. Using the Lerner index we find an increase in competition duringthe privatisation period (1999−2002). This was followed by a decrease in our measure of competition in2003−2004 and a slight revival in 2005. This may seem a surprising finding, as one may have expected that themassive entry of foreign investors into the Czech banking industry would have contributed to enhancing thedegree of banking competition. On the other hand, the decrease in competition or increase in the Lerner indexin 2002−2005 might have been due to the fact that after 2002 banks also offered some riskier and pricierproducts.

Our analysis of the relationship and causality between our measure of competition and estimated cost efficiencysuggests that competition has negatively affected cost efficiency in the Czech banking sector. Although it mayappear counterintuitive, this finding is in accordance with the part of the literature which supports the existenceof a negative link between competition and efficiency. It can be explained by the fact that increased competitionleads to greater monitoring costs for banks (economies of scale mean that a higher number of banks leads tohigher costs) and a reduction of the length of the customer relationship between the bank and the borrower,which reduces efficiency.

Our findings have potentially important implications, as they reveal that efforts to increase competition might bebalanced by a decrease in the cost efficiency of banks, which could result in higher loan rates.

REFERENCES

ALLEN, F., GALE, D. (2004): Competition and Financial Stability, Journal of Money, Credit and Banking, Vol. 36,No. 3, pp. 453−480

ANGELINI, P., CETORELLI, N. (2003): Bank Competition and Regulatory Reform: The Case of the Italian BankingIndustry, Journal of Money, Credit and Banking, Vol. 35, pp. 663−684

ARRELANO, M., BOND, S. (1991): Some Tests of Specification for Panel Data: Monte Carlo Evidence and anApplication to Employment Equations, Review of Economic Studies, Vol. 58, pp. 277−297

BAIN, J. (1951): Relation of Profit Rate to Industry Concentration, Quarterly Journal of Economics, Vol. 65, pp. 293−324

BERGER, A. (1995): The Profit-Structure Relationship in Banking − Tests of Market-Power and Efficient-StructureHypotheses, Journal of Money, Credit, and Banking, Vol. 27, pp. 404−431

BERGER, A., DEYOUNG, R. (1997): Problem Loans and Cost Efficiency in Commercial Banks, Journal of Bankingand Finance, Vol. 21, pp. 849−870

BIKKER, J., HAAF, K. (2002): Competition, Concentration and their Relationship: An Empirical Analysis of theBanking Industry, Journal of Banking and Finance, Vol. 26, pp. 2191−2214

CLAESSENS, S., LAEVEN, L. (2004): What Drives Bank Competition? Some International Evidence, Journal ofMoney, Credit and Banking, Vol. 36, No. 3, pp. 563−583

DEMSETZ, H. (1973): Industry Structure, Market Rivalry and Public Policy, Journal of Law and Economics, Vol. 16,pp. 1−9

DIAMOND, D. (1984): Financial Intermediation and Delegated Monitoring, Review of Economic Studies, Vol. 51,pp. 393−414

FERNANDEZ DE GUEVARA, J., MAUDOS, J., PEREZ, F. (2005): Market Power in European Banking Sectors, Journalof Financial Services Research, Vol. 27, No. 2, pp. 109−137


FRIES, S., TACI, A. (2005): Cost Efficiency of Banks in Transition: Evidence from 289 Banks in 15 Post-CommunistCountries, Journal of Banking and Finance, Vol. 29, pp. 55−81

GELOS, R., ROLDOS J. (2004): "Consolidation and Market Structure in Emerging Market Banking Systems."Emerging Markets Review 5, 1, 39−59

GOLDBERG, L., RAI, A. (1996): The Structure-Performance Relationship in European Banking, Journal of Bankingand Finance, Vol. 20, pp. 745−771

HASAN, I., MARTON, K. (2003): Development and Efficiency of the Banking Sector in a Transitional Economy,Journal of Banking and Finance, Vol. 27, No. 12, pp. 2249−2271

HICKS, J. (1935): The Theory of Monopoly, Econometrica, Vol. 3, pp. 1−20

PODPIERA, A., PODPIERA, J. (2005): Deteriorating Cost Efficiency in Commercial Banks Signals an Increasing Riskof Failure, Working Paper No. 6, Czech National Bank

REININGER, T., SCHARDAX, F., SUMMER, M. (2002): Financial System Transition in Central Europe: The FirstDecade, SUERF Studies No. 16

WEILL, L. (2004): On the Relationship between Competition and Efficiency in the EU Banking Sectors, Kredit undKapital, Vol. 37, No. 3, pp. 329−352

94 OPERATIONAL RISK AND ITS IMPACTS ON FINANCIAL STABILITY

OPERATIONAL RISK AND ITS IMPACTS ON FINANCIAL STABILITY

Věra Mazánková and Michal Němec, CNB

This article illustrates the nature and significance of operational risk with regard to financial stability, using specificexamples of major operational risk events in the Czech Republic and in other countries, and identifies the impactsof the newly introduced capital regulation of operational risk. It looks in detail at the incipient practice of risk-sensitive operational risk measurement using methods based on financial institutions' internal models. We alsoexplore the issue of calculating the mandatory capital coverage for unexpected operational risk losses usinggroup-wide models.

1. INTRODUCTION

January 2008 saw the introduction of a new obligation on all banks and credit unions and selected investmentfirms in the Czech Republic to ensure adequate capital coverage for operational risk in addition to credit risk andmarket risk. This requirement stemmed from the incorporation of Basel II − the new capital framework of the BaselCommittee on Banking Supervision (BCBS) − into the European legislation and subsequently into Czech law.

This article sets out to clarify the nature, aims and practical forms of operational risk management and regulationand the implications for financial stability. In the individual sections we therefore describe:

- the key concepts, starting with the definition of operational risk and a description of its links to the principalfinancial risks, i.e. credit risk and market risk (section 2);

- the prudential requirements in the operational risk area and selected impacts thereof (section 3);- the advanced approaches for operational risk measurement, emphasising the elements thereof having potential

impacts on banks' capital adequacy and thus on their financial condition and stability, focusing specifically onthe mechanisms for allocating capital to subsidiaries in the Czech Republic and on insurance as an importanttechnique for mitigating operational risk in the banking sector (section 4).

2. OPERATIONAL RISK

Just as in any other field of business, factors such as people, internal processes, technological systems and externalevents play a key role in financial institutions. It is in the natural interests of every financial institution to ensurethat such factors provide it with maximum support in achieving its business goals. However, these factorsinherently involve various risks stemming from potential failures which can affect the operations and thus also theoutputs and results of a financial institution.

This leads to the widely accepted definition of operational risk in the banking sector: "Operational risk isdefined as the risk of loss resulting from inadequate or failed internal processes, people and systemsor from external events, including legal risk". This definition, confirmed by best practices,145 is usuallyfollowed immediately by an explicit statement that operational risk excludes strategic and reputational risk.

As is evident from its definition, operational risk, unlike the key financial risks (credit risk and market risk), is notlinked primarily with a financial institution's portfolios (credit, trading, investment), but instead relates to itsprocesses and operations and the main elements thereof − people, systems and technology.146

145 Seven basic types of loss events, or areas of occurrence of operational risk, have been pinpointed from previous best practices: (1) internalfraud, (2) external fraud, (3) employment practices and workplace safety, (4) clients, products and business practices, (5) damage to physicalassets, (6) business disruption and system failures, (7) execution, delivery (including outsourcing) and process management.

146 Under the regulatory rules, events having the character of operational risk associated with market or credit activities, such as credit fraud,exceeding of trading limits, legal shortcomings in the contractual guarantees of receivables, damage resulting from professional shortcomingsin the preparation of new products, programme deficiencies, valuation errors, etc., are also deemed part of operational risk monitoring.

OPERATIONAL RISK AND ITS IMPACTS ON FINANCIAL STABILITY 95

Owing mainly to events such as 9/11, the collapse of energy giant Enron, the rogue trading of a "lone wolf"currency trader at a U.S. subsidiary of Allied Irish Banks and the subsequent failure of that subsidiary, and therecent events at Société Générale, it is no exaggeration to say that literally everyone now has a basic, or at leastintuitive, awareness of operational risk in the financial sector. It is also clear that operational risk events can alsosignificantly affect the reputation, risk profile and financial standing of an institution, as illustrated by theexamples of operational risk events given in the following table.

The growing media coverage of the events mentioned here and of similar events has also encouraged morecomprehensive awareness and analysis. This has gradually led to a higher degree of systemisation of theoperational risk management approaches used by businesses, their regulators and supervisors and by otherinstitutions specialising in financial system soundness and stability. The causes and nature of operational riskevents, however, have a history as long as the finance business itself. The well-known tools and processes widelyused in practice to prevent or limit operational risk events and mitigate their impacts also remain similar.Examples include the segregation and restriction of decision-making and executive processes and powers,screening of fitness for certain professions, various checks and balances, security and management of access toinformation and other assets, mandatory testing and backing-up, contingency and crisis planning, as well as thecreation of budgetary and other internal provisions for operational losses and various types of insurance.

So, as far as the rather abstract concept of operational risk is concerned, we are talking not about a "newlydiscovered" risk, but rather about its newly emerging manifestations − for example operational risk associatedwith outsourcing or with the electronic distribution of financial products and services. We can also say that higherquality (more systematic and comprehensive) and more sophisticated operational risk management methods arenow being applied. Financial institutions are gradually introducing integrated operational risk managementsystems that conform to best practices. At the same time such systems are helping these institutions to complywith the recently introduced mandatory regulation of operational risk.

Table 1 − Selected operational risk events around the world and in the Czech Republic

Cause - Event (Institution) Impact1) / Year

Cheque fraud (group of U.S. retail banks) Failure to ensure segregation of operations − fraud (Barings) Insider trading (Merrill Lynch) Inadequate trading limits and controls (Nomura Securities) Misuse of client accounts by bank employees (ABN AMRO) "Computer" fraud by employees (WGZ Bank) Credit fraud by client − forgery of loan documents (Citibank) Auction system failure (Ebay) Terrorist attack on World Trade Center (WTC) Rogue trading (Société Générale)CR3) − Credit fraud - B.C.L. case (KB)CR − Non-compliance with dealing procedures (ČSOB)CR − Floods (numerous financial institutions)CR − Sporoservis failure − credit fraud (ČS)CR − Cash theft at agency providing services mainly to financial institutionsCR − Fee rounding errors in IT system (KB)

1) Amounts converted at exchange rate valid at time of occurrence or discovery of event.2) Published estimates of impacts vary.3) Sources of information on events in Czech Republic (in same order as in table): Hospodářské noviny (HN) 27 February 2008, HN 6 December 2001, Mladá fronta

(MF) 24 February 2003, MF 5 April 2006, HN 4 December 2007, Euro 28 January 2008.4) Figure for Czech Republic as a whole, institutions and households.Sources of information on events outside Czech Republic: Operational Risk Magazine, Risk Magazine, Incisive Media Ltd., UK.

$12,000 m / 1993$1,600 m / 1995

$100 m / 1997$48,000 m / 1998

$140 m / 1998$200 m / 1998$30 m / 1999

$5,000 m / 19992) / 2001

$7,300 m / 2008up to $180 m / 1999

$35 m / 20014) $2,100 m / 2002

$40 m / 2006$30 m / 2007$10 m / 2007


3. REGULATION OF OPERATIONAL RISK

3.1. Operational risk management

As with other risks, managing operational risk involves defining an institution's overall approach to the risk andoperating appropriate internal systems and processes. These systems and processes should ensure that operationalrisk is identified, assessed, monitored, declared and controlled and/or mitigated. Operational risk mitigationtechniques include ensuring adequate capital coverage for unexpected losses (see section 3.2.) and alternativelytaking out commercial insurance.

From the operational risk management perspective, it is desirable to pay attention − among other things − tooperational risk events that do not have a direct financial impact. For example, the attempt by two Czechbusinessmen to commit a credit fraud with a calculated potential impact of $3,500 million against a well-knownSwiss bank (2006)147 illustrates just how well-founded this requirement is.

The following four key principles are widely regarded as essential for a sound and effective operational riskmanagement system:148

- the creation and development of an appropriate permanent environment (framework) for the systematicmanagement of operational risk;

- the introduction and application of efficient and effective firm-wide and specialised operational risk processesand tools, including adequate capital coverage for unexpected losses due to operational risk;

- independent internal and external review and assessment of operational risk management;- transparency, i.e. disclosure of information on operational risk and operational risk management.

3.2. Capital regulation of operational risk

As mentioned earlier, one of the key operational risk management tools is to maintain adequate capital coveragefor unexpected losses due to operational risk. Apart from recommended148 individual operational risk capitalcoverage, operational risk now ranks alongside credit risk and market risk as one of the three risks with mandatorycapital regulation. The former (Basel I) and current (Basel II) breakdown of regulated banking risks is illustrated bythe following diagram.

147 Hospodářské noviny, 7 February 2008148 For more details, see, for example, Sound Practices for the Management and Supervision of Operational Risk, Basel Committee for Banking

Supervision (2003).

Basel I

Basel II

market risk

credit risk

mandatory capital coverage − global regulation

operationalrisk

otherrisks

otherrisks

individual capitalcoverage

no global capital regulation

market risk

credit risk

mandatory capital coverage − global regulation

" t o t a l r i s k s "

" t o t a l r i s k s "


It is useful at this point to recall some of the basic requirements and aims of Basel II:

- one aim is to raise the quality of systems, processes and methods for risk measurement and managementgenerally, leading eventually also to a reduction of the previous capital charges for credit and in some casesalso for market risk, so that the total capital charge does not generally increase after incorporation of the newseparate operational risk capital charge;

- institutions can choose from among at least three different approaches for determining operational risk capitalcharges, with the simultaneous use (combination) of various approaches permitted under certain conditions;

- institutions are not allowed to revert from a more advanced approach for determining operational risk capitalcharges to a simpler approach without good reason (in order to prevent any attempts at capital arbitrage);

- according to surveys (mostly conducted by the Basel Committee on Banking Supervision), the ratio of the newoperational risk capital charge to the total capital charge is estimated at around 10−15% globally in the longterm.

The impacts of Basel II on the Czech banking sector have also been estimated. According to the QIS 5 surveyconducted in 2005, the ratio of the new operational risk capital charge to the total capital charges of banks inthe Czech Republic was expected to be around 8%.

Real data on operational risk capital charges for the entire Czech banking sector are available from the start of2008. The specific impact of the new mandatory operational risk capital charge on total capital charges is shownin the following chart.

The ratio of the operational risk capital charges to the total capital charges for the Czech banking sector as awhole was 10% at the start of 2008. This figure is broadly in line with the long-term estimates and predictionsof the Basel Committee on Banking Supervision. So, compared to the surveys of the predicted impacts of Basel IIin the Czech Republic conducted in 2005, the ratio of the operational risk capital charges to the total capitalcharges for the sector as a whole is around 2 percentage points higher (up from 8% to 10%). This increase isprobably due mostly to an increase in the proportion of banks using the simplest − and hence the more capital-intensive − approach for determining the operational risk capital charge as compared to the proportion of suchbanks in the survey conducted in 2005.

0

2

4

6

8

10

12

14

0% - 4% 5% - 9% 10% - 14% 15% - 19% 20% - 24% 25% -30% > 30%

Ratio of OR CC to total CC of bank

No.

of b

anks

Chart 1 − Ratios of operational risk capital charge (OR CC) to total capital charges for the Czech banking sector as of 31 January 2008 (sector average: 10%)

Source: CNB, 2008


Specifically, banks in the Czech Republic have the following categories of approaches available for determiningtheir minimum operational risk capital charges:

- the Basic Indicator Approach (BIA)- the Standardised Approach (TSA) or the Alternative Standardised Approach (ASA) - the Advanced Measurement Approaches (AMA).

The first two approaches are based on the assumption that the size of operational risk (the exposure of theinstitution to operational risk) is directly proportional to the value of a particular indicator. The indicator isuniformly defined and based either on income (the BIA and TSA149) or on the volume of loans provided (the ASA).The capital charge corresponds to a fixed percentage of the value of the indicator (12, 15 or 18%). These simplerapproaches are fully standardised, predefined and laid down bindingly and uniformly in the Czech Republic in alegal rule,150 so they are not described in any further detail here.

The most complex and sophisticated approaches are those based on internal models (the advanced approaches,or AMA), which will be examined in more detail in the following section. Generally, the parameters of the AMAreflect the most advanced operational risk management and measurement practices and in this sense offer aguide for all banks.

4. THE ADVANCED APPROACHES TO OPERATIONAL RISK MEASUREMENT

The fundamental quantitative requirement for an AMA operational risk measurement system is that it must havethe following elements: internal data, external data, scenario analysis and business environment and internalcontrol factors. However, more detailed rules − defining, for example, how these mandatory elements should beincorporated into the AMA, how they should be combined and what weights they should have in the overallmeasurement of operational risk − are not laid down in the regulations.

Qualitative requirements are also defined for the AMA. Financial institutions' operational risk managementsystems and processes must comply with explicit "advanced" requirements. These requirements are in principle inline with internationally recognised best practices for operational risk management.

The most valued feature of the AMA is that the quantitative requirements (the main ones are described in moredetail below) are relatively general, which means that institutions can use their tried and tested operational riskmeasurement methods. On the other hand, the generality and flexibility of the regulatory requirements maysimultaneously be the biggest barrier to the use of the AMA, especially for smaller or less sophisticatedinstitutions, as tried and tested methods that can be applied directly under the AMA have not been fullyintroduced yet.

Another feature of the AMA which is also appreciated in practice is that the regulations explicitly allow it to bedeveloped and used on a group-wide basis, i.e. a common model can be created for an entire consolidated group,enabling the group to take advantage of the effects of spreading its operational risks. Consequently, the groupmodel can be used for calculating the operational risk capital charge both for the group as a whole and separatelyfor the individual members of the group. However, an allocation mechanism − whereby the capital charge isderived from the group calculation − is much more frequently used to determine the capital charges for theindividual members of a group. As this approach is easily the most prevalent in the Czech Republic, it will bedescribed in more detail in section 4.3.

The last specific feature of the AMA to be addressed in more detail in this article is the option of using operationalrisk mitigation techniques, most notably insurance, to reduce operational risk capital charges.

149 The standardised approach for calculating the operational risk capital charge is based on the assumption that operational risk exposureexpressed by means of a universal income-based indicator depends additionally on the nature of an institution's business activities. For thispurpose, bank's business activities are divided into eight business lines and three categories with a rising level of operational risk (the riskweight, referred to as beta, is given in parentheses): (1) retail brokerage, retail banking and asset management (12%), (2) commercial banking and agency services (15%), (3) corporate finance, trading and sales, and payment and settlement (18%).

150 Decree No. 123/2007 Coll.


4.1. Elements of the AMA

Internal dataComprehensive information on individual operational risk events (in particular realised losses) is the cornerstoneof operational risk management, and so all institutions should have such an overview no matter what approachthey use to calculate their operational risk capital charges. If an AMA is used, this overview must include anappropriate valuation of the size of the losses ensuing from such events.

The most frequently used data collection system is based on correspondents, i.e. workers who are responsible,among other things, for collecting information on internal operational risk events and recording them in therelevant operational risk system or database. The information obtained is then used for further data analysis and,where relevant, for assessing specific events. This system, based on the collection of information "in the field", ismore difficult to organise and more costly and often fails to ensure that data of the required completeness arecollected. On the other hand, the advantages of this system are that it allows all event-related costs to be takeninto account more accurately and it enables the institution to acquire and assess more operational risk informationand events, including potential losses, indirect costs and events that do not lead directly to operational risk losses.

Another fairly frequently used data collection system is based on the use and analysis of accounting records. Forthis approach, the first step is to select the accounts on which operational risk events are or can be recorded. Theseaccounts are then periodically analysed and any operational risk-related changes on them are transferred to anoperational risk event database. The advantages of this approach are that it ensures more complete collection ofevents with an accounting impact, provided that due care is taken during the initial selection of the trackedaccounts, and the data collection process is less costly. The weaknesses of this system compared to the previousapproach are that a time lag may arise between the occurrence of an operational risk event and the date it isrecorded in the accounts, there is a smaller amount of accompanying analytical information on individual events,and it involves the exclusive use of book valuation of individual events. This valuation method can be tooinaccurate for operational risk management purposes, for example for events related to long term assets wherethe accounting depreciation does not reflect the true value of the asset in question.

With regard to the completeness, accuracy and timeliness of the information recorded, preference is given incurrent practice, including in the Czech Republic, to data collection systems based on a limited number ofcorrespondents with subsequent checks of the completeness of recorded events with data in the accountingsystem; the size of the loss may differ for the reasons mentioned above (in such case the difference should beexplained).

External dataExternal data are included in the AMA primarily to provide additional information on significant, yet infrequentoperational risk events. The data are obtained, for example, from other institutions via membership in aconsortium of institutions that pool information on internal operational risk events, via commercial databases orvia internal event monitoring using the press or other public information sources (owing to increasing mediacoverage, the probability of catching a loss event rises with loss size).

One of the main ways of using external data in the AMA is to incorporate them directly into the internal data. Withthis approach, it is essential to ensure that the external data do not unduly skew the internal data distribution. Suchskewness can have several different causes; for example, consortium or commercial databases only contain lossesexceeding a certain threshold, and this threshold is usually higher than the one used for internal data. One possiblesolution to this problem is to compare the shape of the distribution of the internal losses in a risk category that canreasonably be expected to contain very severe losses with the external data distribution, and on the basis of thatcomparison eliminate the skewness of the external data in other risk categories.

Another very frequently used solution is to apply an appropriate scaling factor to the external data, i.e. to adjustthe amount of the loss recorded by an external institution according to a factor available for one's own institutionand an external one.151 However, one should proceed with caution when scaling, because not every event type or

151 The indicators used for scaling include the ratio of total assets and the ratio of number of employees.

loss amount depends on such factors; for example, losses associated with human error in financial market tradingor certain losses resulting from legal disputes are not necessarily related to institution size.

Scenario analysisThe incorporation of scenario analysis serves a similar purpose as that of external data. These two mandatoryelements are meant to ensure that the AMA captures extraordinary events with very severe losses, events which− given the limited internal data time series used for modelling purposes − may not be recorded among theinternal data. Unlike internal and external data, scenario analysis takes into account expert opinion regarding thepotential future evolution of operational risk. As a result, it is possible to incorporate into the AMA potential"new" losses and other projected trends in an institution's operational risk exposure.

Given the aforementioned similarity of the reasons for including external data and scenario analysis, these twoelements are often combined and incorporated jointly into the AMA. The classic example of this approach is thederivation of potential severe impacts in scenario analyses, where available external data are provided to theexperts as a source of inspiration in their assessment. Even in this case, however, the experts must not ignorepotentially severe loss types specific to the given institution.

Business environment and internal control factorsThe last mandatory element of the AMA consists of methods for incorporating various changes in the business orinternal control environment into the measurement. These methods allow an institution to adjust the capitalcharge calculated on the basis of the previous elements and thereby eliminate the shortcomings inherent primarilyin the internal data (i.e. the assumption that past experience is the best tool for estimating future losses). Byincorporating such factors, the capital charge can be reduced if, for instance, new control mechanisms areintroduced which have a provable impact on the institution's risk profile. Or, conversely, the charge will need tobe increased if, for example, there is significant increase in the institution's activities or it commences newoperations or launches new products. This mandatory element, like scenario analysis, is meant to make a bank'srisk measurements more forward-looking and allow it to take account of changes in qualitative factors.

For this purpose, key risk indicators are used most often in practice. They allow a bank to estimate the future levelof the risks it undertakes. Examples include the number of particular transactions processed by a single employeeand the number of open legal disputes. Although large sets of such indicators are mentioned in the literature, thenumber of risk indicators chosen as key indicators is usually in single or, at most, double figures. Determining thespecific set of key risk indicators appropriate for a particular financial institution is therefore quite a difficultprocess and needs to take account of the institution's specific situation. Another potential tool for assessingchanges in the control environment is the application of risk self-assessments, which involve contacting individualprocess owners tasked with identifying specific process risks, evaluating the adequacy of the existing controls andassessing the residual risks, i.e. those not captured by the controls already in place.

Both tools (risk indicators and risk self-assessments) require "backtesting" to assess the key indicators' risk-prediction ability or the experts' estimation accuracy in risk self-assessments. The predictions are also comparedagainst the internal losses actually realised. Significant deviations should be recorded and explained and, whererelevant, the use of these tools should be modified so as to ensure better agreement in future backtesting.

4.2. Combinations of mandatory AMA elements

Internal data-based AMA − LDA model As for the methodology applied to calculate capital charges using the aforementioned AMA elements, thedominant approach currently used in the Czech Republic and elsewhere is based on the tracking of internaloperational risk events and the subsequent derivation of a mathematical apparatus based on those internal data.This approach is known as LDA.152

If we have access to sufficient internal loss information of the required quality, we can create a model forestimating the total loss. Since, however, individual operational risk events are highly diverse, we first need tocreate a set of homogeneous data that can be expected to be based on the same statistical distribution. Although


152 Loss distribution approach.


this is not specifically required by the regulations, the data are often distributed into risk classes corresponding tothe combination of seven event types145 and eight business lines149 (i.e. 56 risk classes in all). In this case, the eventswithin each of these classes are regarded as events generated from the same distribution and less strictrequirements are usually imposed on the statistical tests of the homogeneity of the data. However, since individualinstitutions (from the economic perspective fortunately) do not have enough data for statistical modellingpurposes, selected risk classes are merged. For such mergers, tests have to be conducted to determine thehomogeneity of the data in each risk class. In practice, one often encounters the solution where the data are splitinto categories according to event type and further broken down by business line for any categories containing asufficient number of observations.

The estimation of the total loss in each risk class is given by the sum of independent, equally distributed randomvariables representing the individual loss amount. As the number of events is also a random variable, we are nottalking about a deterministically determined number of summands. We are talking about a random sum wherethe number of summands corresponds to the realisation of a random variable with a discrete distribution. At thesame time, independence between the number of events and the individual loss amounts is assumed.

A quantile (usually at the 99.9% level) is determined from the total loss distribution. This quantile forms the basisfor the calculation of the operational risk capital charge.

In some cases this problem is not easy to solve, because it is difficult to fit the observed data using a theoreticaldistribution function matching the observed values sufficiently accurately across the whole range of losses.Therefore, combinations of several distributions or, for distributions of large losses, functions based on the theoryof extreme values153 are also used. Since the derivation of the theoretical distribution function of the randomvariable of the total loss is associated with various problems, Monte Carlo simulation is usually employed.

Given the aforementioned dominance of "group" AMAs, this part of the AMA is usually conducted only in parentinstitutions, i.e. outside the Czech Republic, and since the theoretical underpinnings of the models, including themethods for selecting suitable distribution functions and related tests and for calculating the total capital chargeacross the individual risk cells, are adequately described in various sources (see, for example, Cruz, 2002, andMoscadelli, 2004), no further space will be devoted to this issue here.

The aforementioned model incorporates internal data only. The other mandatory elements can be incorporatedinto the approach in several ways, but the authors have not noted any dominant approach that it would be usefulto describe in this article.

Scenario analysis-based AMA − SBA modelTo the best of the authors' knowledge, the second most common approach is one based on scenario analysis,154

although, as emphasised above, it is possible to apply approaches based on other methodologies.

This approach, unlike the scenario analysis used in the LDA method, employs far more scenarios with the aim ofcovering lower risk events as well. Individual experts therefore estimate first the distribution functions of theseverity and frequency of the losses and then the parameters of the chosen distributions. Internal data can thenbe used, for example, to test whether the scenarios created in an area where there is sufficient internal datacorrespond to the scenario analysis estimates.

4.3. Capital allocation using the group AMA approach − potential impacts on the Czech financial sector

All institutions in the Czech Republic that are using or planning to use an AMA in the near future are part of largeinternational financial institutions. The parent companies, as mentioned earlier, are predominantly developing thegroup-wide approach. Individual group members thus contribute all the required data to the model, and thesedata are used to calculate the capital charge at group level, including group diversification effects. This capitalcharge is then distributed across the individual institutions using allocation algorithms often based on readilyavailable indicators such as total assets, gross profit or number of employees. If, however, the diversification

153 The POT (peaks over threshold) approach.154 The SBA (scenario based approach).


effects implicitly included in the group calculation are not removed prior to allocation, it is necessary to verifywhether the allocated capital is commensurate with the size of the risk undertaken by the specific subsidiary (heremeaning a subsidiary bank operating in the Czech Republic, although the information given below is universallyapplicable). For the Czech banking sector, this problem is particularly important, because the vast majority ofbanks are subsidiaries of major European banks, some of which already use an AMA or are preparing intensivelyto do so.

The problem of allocation associated with diversification effects can be illustrated using the following example,which sets out to compare the individually calculated capital charge with the result derived from the groupcalculation155 with subsequent application of the now common allocation mechanism with no adjustment fordiversification effects.

For the purposes of the simulation, the frequently employed distribution functions were chosen − the Poissondistribution for the distribution of the number of events and the log-normal distribution for the distribution of theindividual loss amounts. For the sake of simplicity, we assumed that the group consists of three identicalinstitutions. Hence, the same distribution functions describing the operational risk level were chosen for both theindividual and group calculation, with relevant adjustment of the parameters in the case of the determination ofthe number of losses. For the same reason, an allocation ratio of one-third of the group capital charge was setfor the allocation of the group capital charge to each of the institutions, in line with simple allocation mechanisms.

The specific parameters used in the simulation were the following:

Log-normal distribution (loss amount)156 parameter µ = 10parameter σ = 2157

Poisson distribution (number of events): parameter λ = 5 (individual calculation)158

parameter λ = 15 (group calculation).

Using the chosen parameters, a Monte-Carlo simulation was performed separately for the group and individualcalculation. In the simulation the following series of steps was followed in both cases:

1) the total loss amount for the period was simulated 100,000 times, and in each stepa) the number of events (the random variable from the Poisson distribution with the relevant parameter)

was generated,b) according to the value of the random variable from the previous step, the corresponding number of

individual losses from the log-normal distribution was generated,c) the total loss in this step was determined,159

2) the quantile was calculated from the total losses generated in step 1 at a confidence level of 99.9%, which corresponds to the regulatory requirement for the operational risk capital charge,

3) for the simulation of the group calculation, the resulting quantile was multiplied by 1/3 as explained above, due to the subsequent use of the allocation mechanism.

The outcome of the above simulation is that an institution which is not a group member or which performs thecalculation itself would have to maintain operational risk capital coverage of around 26 million currency units inthe case of the individual calculation, while an institution incorporated into the group calculation with subsequentallocation would be able to maintain capital of just 16.5 million currency units thanks to group diversificationeffects.

155 The calculation performed at group level using data from the individual group members.156 These parameter values are only illustrative, but similar values are also used in practice.157 The mean of this random variable is equal to exp(µ+ σ2/2) and the variance exp(2µ+ σ2) exp(σ2-1).158 The mean and variance of this random variable are equal to parameter λ.159 In terms of symbols , where S is the total loss amount for the period, Y is the random number of events for the period and Xi is the

random amount of one loss.

In order to reduce the statistical error of adopting a conclusion based on just one simulation, the aforementionedsimulation was repeated 100 times. The result is shown in the chart.

The repetition of the simulation confirmed the validity of the above conclusion that the group-wide calculationwith subsequent allocation leads to a significantly lower capital charge than the individual calculation. It is alsoworth mentioning that according to the results presented in the chart even the highest capital charge calculatedusing the allocation mechanism is 25% lower than the lowest capital charge calculated on an individual basis.

Since parent companies do not usually provide group members with any legally binding guarantees to provideadditional capital where necessary to cover losses due to operational risk, this capital saving within the group asa whole is not sufficiently justifiable from the subsidiary's perspective. It is therefore necessary to perform tests ofthe adequacy of the capital allocated from the point of view of the individual group members, especially whenallocation mechanisms are applied that do not eliminate group diversification effects.

Clearly this is just an illustrative example and the capital saving may be different in practice. However, the exampledemonstrates that if group diversification effects are not eliminated prior to applying the allocation mechanism,the capital charge can be significantly underestimated from an individual perspective for individual subsidiaries.One of the specific responses of the CNB supervisory authority to this fact has been to set a prudential benchmark(threshold) for the operational risk capital charge.160

4.4. Insurance as an eligible technique for reducing the operational risk capital charge

When using the AMA, unlike all the other available approaches for calculating operational risk capital charges, abank is allowed to reduce its calculated capital charge if some of its operational risks are insured and transferredto insurers outside its financial group, provided that the insurance policies meet other specified conditions. In suchcase, the operational risk capital charge can be reduced by up to 20%. Besides insurance, other operational riskmitigating techniques can be applied, although their use for reducing capital charges is conditional on priorsupervisory assessment and approval. The application of derivatives is most often mentioned in this context, butthis option is little used in practice as yet.


160 Official Information of the Czech National Bank of 16 November 2007 regarding the prudential rules for banks, credit unions and investmentfirms − a benchmark for the operational risk capital charge.

0

5,000,000

10,000,000

15,000,000

20,000,000

25,000,000

30,000,000

35,000,000

0 20 40 60 80 100

Cap

ital c

harg

ein

giv

en c

urre

ncy

Individual calculation Allocated capital

Chart 2 − Simulated OR capital charge

Source: CNB

In the Czech Republic, too, insurance mitigation is the only such technique being considered at present. Its impactis so far relatively insignificant and the savings being achieved are only a few per cent of the operational riskcapital charge. Although the technique only involves the transfer of risk out of the relevant institution or group,and not out of the financial sector, this issue is not currently a priority with regard to the financial sector as awhole. However, it requires continued vigilance.161

5. CONCLUSIONS

The main results of the study include an initial assessment of the real impacts of the newly introduced mandatoryoperational risk capital charge on the total capitalisation (capital adequacy) of banks in the Czech Republic and acomparison with earlier estimates and predictions. Importantly, we have demonstrated that the operational riskcapital charge is potentially underestimated where a group-wide operational risk measurement model is used andcapital is subsequently allocated to banks in the Czech Republic. Other important findings include confirmationof the potential of operational risk to significantly affect financial institutions' risk profiles, as well as the limitedscope for limiting the impacts of operational risk on the financial sector as a whole given the dominant status ofinsurance as an eligible technique for mitigating operational risk in the banking sector.

REFERENCES:

ÁLVAREZ, G. (2006): Operational Risk Economic Capital Measurement: Mathematical Models for Analysing LossData, in Davis, E. (ed.): The Advanced Measurement Approach to Operational Risk, ISBN 1 904339 88 3

BCBS (2003): Sound Practices for the Management and Supervision of Operational Risk, BCBS

BCBS (2006): International Convergence of Capital Measurement and Capital Standards: A Revised Framework,BCBS

BCBS (2007): Principles for Home-Host Supervisory Cooperation and Allocation Mechanisms in the Context ofAdvanced Measurement Approaches, BCBS

CEBS (2006): Guidelines on the Implementation, Validation and Assessment of Advanced Measurement (AMA)and Internal Rating Based (IRB) Approaches, CEBS GL 10

Cruz, M. G. (2002): Modeling, Measuring and Hedging Operational Risk, John Wiley & Sons, ISBN 0 471 51560 4

JOINT FORUM (2003): Operational Risk Transfer Across Financial Sectors, Joint Forum(http://www.bis.org/list/jforum/index.htm)

MOSCADELLI, M. (2004): The Modelling of Operational Risk: Experience with the Analysis of the Data Collectedby the Basel Committee, in Davis, E. (ed.): Operational Risk: Practical Approaches to Implementation, ISBN 1 904339 46 8


161 See, for example, Joint Forum (2003).

GLOSSARY 105

GLOSSARY

Asset-Backed Security (ABS)A type of debt security whose yield is secured with cash flows from a set ofunderlying assets (mortgage loans, income from payment cards issued, carpurchase loans, etc.). ABSs backed by payments from mortgage loans onresidential property are called Residential Mortgage-Backed Securities (RMBS).Asset-Backed Commercial Papers (ABCP), which are issued to obtain short-termcredit, are a variant of ABSs.

Balance-sheet liquidity The ability of an institution to meet its obligations in a corresponding volume andterm structure.

Capital adequacy ratio The ratio of regulatory capital to total risk-weighted assets. Tier 1 capital adequacyis the ratio of Tier 1 capital to total risk-weighted assets (see also Tier 1).

Carry trade A speculative strategy on the financial markets where an investor borrows moneyin a currency with a lower interest rate and invests it in a currency with a higherinterest rate in order to realise the profit arising from the interest rate differential(a similar transaction can be carried out using financial derivatives). However, thisprofit can only be realised on the assumption that any movement of the exchangerate between the financing currency and the investment currency does noteliminate the gain from the interest rate differential.

Collateralised debt obligation (CDO)A security backed by debt, with various debt receivables serving as the underlyingportfolio. By contrast with traditional debt-backed securities, several risk variants(tranches) of debt securities, characterised by specific risk profiles and yields, areissued against a single underlying portfolio. Riskier tranches offer higher yields butalso result in higher losses in the case of default in the underlying portfolio. CDSscan also be used as underlying assets; such obligations are called synthetic CDOs.

Conduits Special-purpose vehicles established by banks. Like SIVs, they sell short-termliabilities (commercial papers) and use the funds raised to buy long-term asset-backed assets (mortgage loans, car purchase loans, income from payment cardsissued, etc.). Banks use them to securitise loans provided and to a large extentguarantee their operations.

Credit default swap (CDS) A credit derivative in which the buyer of the collateral undertakes to pay the sellerperiodical fixed payments ("swap premium") for the duration of the contract inexchange for a conditional payment of the counterparty in the case of default ofthe "reference entity" to which the agreement refers. If default does not occur, thecontract terminates at a specified time and the seller only gains a premium fortaking on the potential credit risk.

Default Default is defined as a breach of the debtor's payment discipline. The debtor is indefault at the moment when it is probable that he will not be able to repay hisobligations in a proper and timely manner, without recourse by the creditor tosettlement of the claim from the security, or when at least one repayment (theamount of which deemed by the creditor to be significant) is more than 90 dayspast due.

Default rate The 12-month default rate is the number of new defaulters over a 12-monthreference period as a proportion of the total number of entities existing in thatperiod. The default rate can also be defined analogously in terms of volume basedon the obligations assumed by debtors.

106 GLOSSARY

Equalisation provision The equalisation provision is set aside for individual areas of non-life insurance andis intended to equalise increased insurance claim costs arising due to fluctuationsin loss ratios as a result of facts independent of the will of the insurance company.

Flight to quality A situation on the financial markets where investors sell en masse all risky assetsand buy only high-quality government bonds of advanced countries with very lowrisk.

Herfindahl index (HI) The sum of the squares of the market shares of all entities operating on a givenmarket. It expresses the level of concentration in the market. It takes valuesbetween 0 and 10,000. The lower the HI, the less concentrated the market.

Interest rate spread Also interest rate differential; the spread between the interest rate on a contract(deposit, security) and a reference interest rate.

IRB approach Internal Rating Based Approach − an approach to the calculation of the capitalrequirement in relation to the credit risk of the investment portfolio and the risk ofdiffusion of the portfolio, based on internal ratings. It is one of the most importantinnovations under the new capital framework. The IRB approach is laid down inCNB Decree No. 123/2007 Coll., on prudential rules for banks, credit unions andinvestment firms.

Leveraged loans Loans with a high degree of leverage, granted to companies whose external debtexceeds a level that is considered normal. These loans have a higher probability ofdefault and banks charge relatively high interest rates on them. They are oftenused for a specific purpose − typically to acquire a controlling interest in acorporation using borrowed money (leveraged buy-out, LBO).

Liquidity Money in the broader sense (cash, short-term assets quickly exchangeable forcash, etc.).

Loan-to-value (LTV) ratio The ratio of a loan to the value of pledged property.

Market liquidity The ability of market participants to carry out financial transactions in assets of agiven volume without causing a pronounced change in their prices.

Monoline insurers Specialised institutions insuring against the risk of default on bond issues andsecuring high ratings for them (originated primarily to insure municipal bondissues). This relatively small sector attracted increased attention during the creditcrisis in 2008, as it also started insuring credit derivatives, including those createdthrough the securitisation of subprime mortgages, in a search for yield.

Natural population increaseThe difference between the number of live births and the number of deaths in thesame period of time in a given area. See also Total population increase.

Originate-and-distribute modelA bank business model under which the bank first provides loans or funds forlending to an intermediary and subsequently sells on these loans by means ofsecuritisation, thereby distributing the original credit risk among other entities.

Over the counter (OTC) operationsOperations not conducted on an organised market.

Present value of a basis point (PVBP)The change in the real value of an instrument given a parallel shift in the interestrate curve of 1 basis point, i.e. 0.01%.

GLOSSARY 107

Price-to-income The ratio of the price of a flat (68 m2) to the average annual wage in a givenregion.

Price-to-rent The ratio of the price of a flat to the annual rent. The price-to-rent ratio is theinverse of the rent return.

Property developers/developmentsCompanies/projects whose aim is to build a complex of residential and commercialproperty. Property developers' work includes choosing an appropriate site, settingup a project, obtaining the necessary permits, building the necessary infrastructure,constructing the buildings and selling the property. Developers also often organisepurchase financing for clients and frequently lease or manage the property once itis built (especially in the case of commercial property). Given the combination ofconstruction activity and speculative property purchases, developers' results arestrongly dependent on movements in property prices.

Property supply prices Property sale supply prices in estate agencies. Supply prices should be higher thantransfer prices. Property supply prices in the Czech Republic are published, forexample, by the CZSO and the Institute for Regional Information (which alsopublishes data on market rent supply prices). See also Property transfer prices.

Property transfer prices Prices based on Ministry of Finance statistics from property transfer tax returns andpublished by the CZSO. These prices are the closest to actual market prices in termsof methodology, but are published with a time delay. See also Property supplyprices.

Rent return The ratio of the annual supply rent to the supply price of the flat. It is the inverseof the price-to-rent indicator.

RMBS See Asset-Backed Security

Securitisation A credit risk transfer method creating a new marketable security from a set ofilliquid assets. The cash flows from the security depend on the cash flows from theunderlying assets. Securitisation takes place through an SPV (Special PurposeVehicle). The original owner transfers the underlying assets to this companyagainst cash payment and the SPV issues new securities based on the underlyingassets (called asset-backed securities, ABSs) and sells them to investors. Credit riskis thus transferred from the original owner of the receivables to ABS holders.

Shadow banking system A relatively new term that started to be used on a larger scale in connection withthe credit crisis in 2007. It is a heterogeneous group of financial institutions thatenable the banking sector to provide more loans with a given level of capital orgrant various forms of loans themselves. According to Wikipedia, for example, theshadow banking system or shadow financial system is largely formed by non-bankfinancial institutions that have short-term and liquid liabilities and long-term andless liquid assets. The system includes SIVs, conduits, money market funds,investment banks, hedge funds, monoline insurers and other non-bank entities.These institutions are subject to market risk, credit risk and especially liquidity riskin respect of the rollover of securities or the loans used to finance them. As theyare not depositary institutions, they do not have access to the central bank'slender-of-last-resort support. In the event of liquidity problems, they could gobankrupt if unable to refinance their short-term liabilities.

Solvency Solvency in the insurance sector is the ability of an insurer to meet its insuranceobligations, i.e. to settle eligible insurance claims arising from insured losses.Solvency II − a new regulatory framework prepared by the European Commission− is a set of rules for European insurance companies and reinsurers laying down

108 GLOSSARY

quantitative requirements, qualitative requirements, prudential rules, compliancewith market discipline and disclosure duties.

Structured Investment Vehicles (SIVs)Special-purpose companies acting as funds, raising money by issuing short-termsecurities (commercial paper), investing it by purchasing long-term securities andmaking profit from the difference between the interest rates on short-termliabilities (asset-backed securities or corporate bonds) and long-term assets. SIVshave an open structure, i.e. they operate continuously, buying new assets as othersmature. By contrast with conduits, they cannot rely directly on obtaining liquidityfrom their parent banks at times of financial turbulence.

Subprime A relatively risky segment of clients with worse expected payment discipline (e.g. clients with a poor credit history, a higher risk of loss of employment, etc.).

Technical provisions Under the Act on Insurance, an insurer must set aside technical provisions to meetinsurance obligations which are either likely to be incurred or certain to be incurredbut uncertain as to amount or as to the date on which they will arise.

Tier 1 The highest quality and, for banks in the Czech Republic, also the most significantpart of regulatory capital. The dominant components of Tier 1 are equity capital,retained earnings and mandatory reserve funds.

Total population increase The sum of the natural population increase and net migration in the same periodof time in a given area. Net migration is the difference between immigration intoand emigration from a given area in the same period of time. See also Naturalpopulation increase.

Value at risk The size of loss, with predefined probability, which a bank may suffer whenholding a current portfolio for a certain period if market factors (e.g. interest rates,exchange rates) develop unfavourably.

Yield spread Also yield differential; the spread between the yield on a bond and the yield on areference ("benchmark") bond.

ABBREVIATIONS 109

ABBREVIATIONS

ABCP asset-backed commercial paperABS asset-backed securitiesAFAM ČR Association of Funds and Asset Management of the Czech RepublicAKAT Czech Capital Market AssociationAPF ČR Association of Pension Funds of the Czech RepublicATM automated teller machineb.p. basis pointČAP Czech Insurance AssociationCAR capital adequacy ratioCCR Central Credit RegisterCDO collateralised debt obligationCDS credit default swapCEBS Committee of European Banking Supervisors CEIOPS Committee of European Insurance and Occupational Pensions SupervisorsCERTIS Czech Express Real Time Interbank Gross Settlement SystemCESR Committee of European Securities RegulatorsČLFA Czech Leasing and Finance AssociationCNB Czech National BankCSD Central Securities DepositoryCZ-NACE Industrial Classification of Economic ActivitiesCZEONIA Czech OverNight Index Average (reference O/N interest rate on the interbank market)CZK Czech korunaCZSO Czech Statistical OfficeEC European CommissionECB European Central BankEEA European Economic AreaEIB European Investment BankEMBI Emerging Market Bond Index EONIA Euro OverNight Index Average (reference O/N interest rate on the interbank market)ESCB European System of Central BanksEU European UnionEU-12 euro area countries before 2007EUR euroEURIBOR Euro InterBank Offered Rate (reference interest rate on the interbank market)FED Federal Reserve System (the US central bank)FRA forward rate agreementGDP gross domestic productHI Herfindahl indexHUF Hungarian forintIBRD International Bank for Reconstruction and DevelopmentIF investment firmIMF International Monetary FundIPB Investiční a Poštovní banka, a. s.IRB Internal Rating Based Approach, an approach within the Basel II framework for capital

adequacy of banks

110 ABBREVIATIONS

IRI Institute for Regional InformationIRS interest rate swapJPY Japanese yenLGD loss given defaultLIBOR London InterBank Offered Rate (reference interest rate on the interbank market)LTV loan-to-value ratioMCR minimum capital requirement − the minimum required capital for calculation of the solvency

of insurance companies and reinsurersMF ČR Ministry of Finance of the Czech RepublicMiFID Markets in Financial Instruments DirectiveMNB Magyar Nemzeti Bank (the Hungarian Central Bank)MRD Ministry for Regional DevelopmentO/N overnightOECD Organisation for Economic Cooperation and DevelopmentOMF open-ended mutual fundOR operational riskOTC over-the-counter p.p. percentage pointPD probability of defaultPLN Polish zlotyPRIBOR Prague InterBank Offered Rate (reference interest rate on the interbank market)PX (PX-Glob) Czech stock market indexQIS quantitative impact studyRMBS residential mortgage-backed securitiesRoA return on assetsRoE return on equityRoS return on salesSCR solvency capital requirement − the minimum solvency requirement for risks undertaken by

insurance companies and reinsurersSIVs structured investment vehiclesSKD Short-Term Bond SystemSKK Slovak korunaSME small and medium-sized enterprisesUSD US dollarVA value addedVAT value added taxWIG Polish stock market index

ABBREVIATIONS 111

AT Austria AU AustraliaBE Belgium BG BulgariaCA CanadaCH SwitzerlandCY Cyprus CZ Czech RepublicDE GermanyDK Denmark EE EstoniaES Spain FI Finland FR France GR GreeceHU HungaryIE Ireland IT Italy JP Japan

KR KoreaLT LithuaniaLU Luxembourg LV LatviaME MexicoMT MaltaNL NetherlandsNO NorwayNZ New ZealandPL PolandPT PortugalRO RomaniaSE SwedenSI SloveniaSK SlovakiaTR TurkeyUK United KingdomUSA United States

Country abbreviations:

112 SELECTED FINANCIAL STABILITY INDICATORS

Selected financial stability indicators

Financial soundness of banks

Capital adequacy (%)Tier 1 capital adequacy (%)Non-performing loans / total gross loans (%)Sectoral breakdown of total loans (%)

- households- sole tradersnon-financial corporations- other (incl. non-residents)

Return on assets (%)Return on equity (%)Quick assets / total assets (%)Quick assets / client deposits (%)Net open position in foreign exchange / capital (%)

Macroeconomic environment

Real GDP growth (year on year, %)Consumer price inflation (end of period, %)Public finance deficit / GDP (%)Public debt / GDP (%)Trade balance / GDP (%)Balance of payments current account / GDP (%)Monetary policy 2W repo rate (end of period, %)

Financial markets

1Y PRIBOR (average, %)10Y government bond yield (average, %)Eurobond spread (EMBI spread, b.p.)CZK/EUR exchange rate (average)Change in the PX stock index (% year on year, end of period)

Real estate market

Total change in residential property prices (transfer prices, % year on year)Change in flat prices (supply prices according to IRI, % year on year)Flat price / average annual wageFlat price / rent

Non-financial corporations

Return on equity (%)Debt (% of total assets)Debt (% of GDP)

- loans from Czech banks (% of GDP)- loans from Czech non-bank financial corporations (% of GDP)- other (incl. financing from abroad, % of GDP)

Interest coverage ratio (earnings / interest expense, %)12M default rate (average, %)

Households (incl. sole traders, excl. 12M default)

Debt / gross disposable income (%)Debt / financial assets (%)Net financial assets (total financial assets − total liabilities, % of GDP)Debt / GDP (%)

- loans from Czech banks to households (% of GDP)- loans from Czech non-bank financial corporations to households (% of GDP)

2005 2006 2007 2008January February March

11.9 11.4 11.5 11.6 11.9 12.311.3 10.0 10.3 11.0 11.2 11.64.1 3.6 2.6 2.7 2.7 2.8

32.2 35.0 37.5 37.7 38.0 38.12.8 2.5 2.2 2.2 2.2 2.2

44.6 44.9 41.7 41.5 41.5 41.520.4 17.5 18.7 18.6 18.4 18.2

1.4 1.2 1.3 1.5 1.4 1.425.2 22.5 24.5 29.8 26.3 26.032.8 30.4 24.0 25.9 24.8 25.350.5 45.5 36.6 39.2 37.1 38.50.1 0.3 0.0 1.9 0.9 0.9

6.4 6.4 6.5 … … …2.2 1.7 5.4 7.5 7.5 7.1

-3.6 -2.7 -1.6 … … …29.7 29.4 28.7 … … …2.0 2.0 3.3 … … …

-1.6 -3.1 -2.5 … … …2.0 2.5 3.5 3.5 3.8 3.8

2.1 2.7 3.4 4.2 4.1 4.23.6 3.8 4.3 4.6 4.6 4.7

17.0 23.0 26.0 32.0 35.0 44.029.8 28.3 27.8 26.1 25.4 25.242.7 7.7 14.2 -9.8 -1.7 -9.4

4.9 2.9 3.96* … … …-2.3 11.3 34.7 … … 28.24.2 4.0 4.1 … … …

15.0 16.9 22.8 … … 23.3

9.5 10.5 10.1 … … …46.5 47.5 48.9 … … …40.9 39.3 41.7 … … …14.8 19.7 20.9 … … …4.7 4.7 4.7 … … …

22.1 15.0 16.1 … … …9.5 11.8 9.8 … … …2.6 2.1 2.8 … … …

34.0 40.3 48.3 … … …22.5 26.0 29.6 … … …

… 41.5 41.1 … … …17.3 20.3 23.4 … … …10.7 15.3 18.8 … … …3.1 3.1 3.6 … … …

SELECTED FINANCIAL STABILITY INDICATORS 113

Selected financial stability indicators −− continued

Households (incl. sole traders, excl. 12M default)

- loans from Czech banks to sole traders (% of GDP)- loans from Czech non-bank financial corporations to

sole traders (% of GDP)- other (incl. financing from abroad, % of GDP)

Interest expenses / gross disposable income (%)12M default rate of households (average, %)

Financial sector

Assets / GDP (%)Bank assets / GDP (%)

Banking sector

Share in financial sector assets (%)Client loans / bank assets (%)Client deposits / client loans (%)Growth in loans (%, end of period, year on year)

totalhouseholds- loans for house purchaseconsumer creditsole tradersnon-financial corporationsloans for house purchase (CZ-NACE 70)

Non-performing loans / total loans (%)householdsloans for house purchaseconsumer creditsole tradersnon-financial corporations

Non-bank financial corporations

Share in financial sector assets (%)Premiums written / GDP (%)Solvency of insurance companies: life insurance (%)Solvency of insurance companies: non-life insurance (%)Change in financial investment of insurance companies (%)Return on equity of insurance companies (%)Claim settlement costs / net technical provisions (life, %)Claim settlement costs / net technical provisions (non-life, %)Change in assets managed by pension funds (%)Return on equity of pension funds (%)Growth in loans from non-bank financial corporations engaged in lending (%)

totalhouseholdsnon-financial corporations

Composite indicators**

Banking stability index (average for period)Creditworthiness index for non-financial corporations (average for period)Market liquidity index (average for period)

2005 2006 2007 2008January February March

0.9 1.1 1.1 … … …

0.4 0.4 0.5 … … …2.2 0.3 0.1 … … …1.1 1.3 1.8 … … …… … 2.9 … … …

134.4 133.0 141.9 … … …98.9 97.5 105.3 … … …

73.8 73.3 74.2 … … …39.5 45.2 48.4 46.8 48.0 48.065.0 66.7 65.6 66.2 66.9 65.8

16.7 19.9 26.4 25.4 24.8 24.534.0 30.4 35.1 34.9 34.5 33.234.1 32.5 37.6 37.5 36.8 35.236.8 26.5 26.1 26.7 27.2 26.616.9 7.7 8.7 8.7 7.1 5.714.3 20.8 17.2 16.4 16.7 16.536.5 37.0 37.4 38.5 37.6 36.8

3.2 2.9 2.7 2.8 2.8 2.81.6 1.6 1.5 1.6 1.6 1.67.8 7.3 6.6 6.7 6.7 6.7

10.7 9.2 7.2 7.6 7.7 7.85.1 4.4 3.1 3.2 3.2 3.4

26.2 26.7 25.8 … … …3.9 3.8 3.7 … … …

325 301 … … … …339 327 … … … …11.6 8.9 8.0 … … …13.5 24.6 21.7 … … …12.1 10.3 12.8 … … …69.4 71.7 61.4 … … …20.9 18.2 14.6 … … …

… 121.8 111.7 … … …

… 7.4 17.7 … … …… 9.2 27.2 … … …… 5.8 11.2 … … …

0.6 0.5 0.3 … … …0.971 0.973 0.972 … … …

0.2 0.3 0.1 -0.3 -0.2 …

* 2007 H1** see Part I of the report for the methodology and interpretation of the composite indicators

Issued by: CZECH NATIONAL BANKNa Příkopě 28115 03 Prague 1CZECH REPUBLIC

Contact:COMMUNICATIONS DEPARTMENTTel.:+420 22441 3494Fax: +420 22441 2179

http://www.cnb.cz

Produced by: Studio PressDesign: Jerome s.r.o.

financial stability report 2007 - cnb.cz · maintaining financial stability is defined as one of...

Documents